Hafmeister, Physics of Societal Issues

Physics of
Societal Issues
Calculations on National Security,
Environment, and Energy
David Hafemeister
P1: OTE/SPH P2: OTE
SVNY342-Hafemeister March 29, 2007 13:54
David Hafemeister
Physics of Societal Issues

Calculations on National Security,
Environment, and Energy
iii
P1: OTE/SPH P2: OTE
David Hafemeister
Department of Physics
California Polytechnic State University,
San Luis Obispo, CA 93407
USA
dhafemei@calpoly.edu
Cover illustration: Cover artwork by Kathryn Bay, Enrico Fermi photo

from AIP Emilio Segre Visual Archives.
Library of Congress Control Number: 2007920188
ISBN-10: 0-387-95560-7 e-ISBN-10: 0-387-68909-5

ISBN-13: 978-0-387-95560-5 e-ISBN-13: 978-0-387-68909-8
Printed on acid-free paper.

C 2007 Springer Science+Business Media, LLC
All rights reserved. This work may not be translated or copied in whole or in part without the written permission
of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except
for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known
or hereafter developed is forbidden.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not
identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary
rights.
9 8 7 6 5 4 3 2 1
springer.com
iv
P1: OTE/SPH P2: OTE
I dedicate this book to my wonderful family of Gina, Andrew, Laurence, Jason, Heidi,
Craig, Matthieu, Adeline, and Alexandre. You have been a joyful sounding board
during our bicoastal dances of legislation and treaties.
There is much about life I do not understand, but at least I remembered this guidepost
from one of my professors:
“The perfect life is one which is guided by knowledge and inspired by love.”
[Professor Paul Schilpp, Northwestern University, 1957]
v
P1: OTE/SPH P2: OTE
Contents
Preface .............................................................................................. xiii
National Security
1 Nuclear Weapons ............................................................................ 3

1.1 Nuclear Age .............................................................................. 3
1.2 Fission Energetics ....................................................................... 5
1.3 Scaling Laws and Critical Masses .................................................. 8
1.4 Efficiency and Neutron Generations .............................................. 11
1.5 Plutonium Implosion Weapons..................................................... 12
1.6 Boosted Primaries and H Bombs ................................................... 14
1.7 Neutron Bomb ........................................................................... 18
1.8 Nuclear Weapon Effects............................................................... 20
2 The Offense: Missiles and War Games............................................... 31

2.1 Rocket Equation ......................................................................... 31
2.2 ICBM Trajectories ....................................................................... 33
2.3 ICBM Accuracy .......................................................................... 34
2.4 GPS Accuracy ............................................................................ 36
2.5 Kill Probability = f (CEP, H, Y, R, n, Fratricide)................................ 38
2.6 Nuclear Conflicts........................................................................ 41
2.7 Conventional Conflicts ................................................................ 50
3 The Defense: ABM/SDI/BMD/NMD ................................................. 55

3.1 ABM History ............................................................................. 55
3.2 Target Interactions ...................................................................... 57
3.3 Nuclear ABMs ........................................................................... 59
3.4 Particle Beam Weapons................................................................ 60
3.5 Laser Weapons ........................................................................... 62
3.6 Orbital Chemical Lasers............................................................... 63
3.7 Earth-Based Lasers ..................................................................... 65
vii
P1: OTE/SPH P2: OTE
viii Contents
3.8 X-ray Laser Pumped with a Nuclear Explosion.............................. 67

3.9 Kinetic Kill Vehicles .................................................................. 70
3.10 Airborne Laser ......................................................................... 72
3.11 AntiSatellite Weapons................................................................ 73
3.12 Rail Guns................................................................................. 74
4 Verification and Arms Control Treaties............................................... 77

4.1 Verification Context................................................................... 77
4.2 Arms Control Treaties................................................................ 78
4.3 Optical Reconnaissance.............................................................. 81
4.4 Adaptive Optics........................................................................ 83
4.5 Digital Image Processing ............................................................ 84
4.6 Infrared Reconnaissance............................................................. 86
4.7 Radar Reconnaissance................................................................ 88
4.8 Nuclear Tests in Space ............................................................... 90
4.9 Atmospheric Nuclear Tests ......................................................... 92
4.10 Underground Nuclear Tests........................................................ 93
4.11 How Much Verification Is Enough?.............................................. 99
5 Nuclear Proliferation ....................................................................... 105

5.1 Proliferation: Baruch to Iraq........................................................ 105
5.2 Uranium Enrichment................................................................. 116
5.3 Separative Work Units ............................................................... 119
5.4 Nonproliferation in the Former USSR........................................... 121
5.5 Plutonium Production ............................................................... 124
5.6 MTCR and Scuds ...................................................................... 128
Environment
6 Air and Water Pollution ................................................................... 137

6.1 Acid Rain pH ........................................................................... 138
6.2 Clean Air Act and Allowance Trading .......................................... 138
6.3 Pollution Scaling....................................................................... 141
6.4 Power Plant Plumes .................................................................. 144
6.5 Automobile Emissions in the LA Air Basin.................................... 148
6.6 Stratospheric Ozone .................................................................. 150
6.7 Purifying Water ........................................................................ 153
6.8 Environmental Chemistry .......................................................... 156
6.9 Flowing Water .......................................................................... 158
7 Nuclear Pollution ............................................................................ 163

7.1 Low-Dose Radiation.................................................................. 163
7.2 Loss-of-Coolant Reactor Accidents............................................... 169
7.3 Plume of 137 Cs from a LOCA....................................................... 173
P1: OTE/SPH P2: OTE
Contents ix
7.4 Weapon Accident Plutonium Plume ........................................... 177

7.5 Dirty Bombs ........................................................................... 180
7.6 Fault Tree Analysis .................................................................. 181
7.7 Geological Repositories ............................................................ 185
7.8 Indoor Radon ......................................................................... 190
8 Climate Change............................................................................. 197

8.1 Introduction ........................................................................... 197
8.2 CO2 Projections ....................................................................... 201
8.3 Upper-Atmospheric and Surface Temperatures ............................ 205
8.4 Temperature Refinements ......................................................... 210
8.5 Link Between CO2 and Temperature........................................... 213
8.6 Solar and Oceanic Variations ..................................................... 216
8.7 Heat Islands ........................................................................... 221
8.8 Policy Options ........................................................................ 224
9 Electromagnetic Fields and Epidemiology ........................................ 233

9.1 Power Line Health Effects? ....................................................... 233
9.2 Epidemiology ......................................................................... 238
Energy
10 The Energy Situation ..................................................................... 249

10.1 Introduction ........................................................................... 249
10.2 Energy Orders-of-Magnitude .................................................... 260
10.3 Fossil Fuel Models ................................................................... 262
10.4 Energy Rates of Change............................................................ 268
10.5 Population and Sustainability.................................................... 269
10.6 Single and Combined Cycle Power Plants ................................... 271
10.7 LNG Explosions ...................................................................... 274
11 Energy in Buildings ....................................................................... 279

11.1 Heat Transfer .......................................................................... 280
11.2 Single- and Double-Glazed Windows ......................................... 283
11.3 Annual Heat Loss.................................................................... 284
11.4 Energy Standards .................................................................... 288
11.5 Scaling Laws for Buildings........................................................ 290
12 Solar Buildings.............................................................................. 299

12.1 Solar Flux............................................................................... 299
12.2 Solar Collectors ....................................................................... 302
12.3 Integrated Solar Flux................................................................ 304
12.4 Solar Hot Water....................................................................... 307
12.5 Active Solar Space Heat............................................................ 309
12.6 Passive Solar Thermal Flywheel................................................. 310
P1: OTE/SPH P2: OTE
x Contents
13 Renewable Energy ......................................................................... 316

13.1 Sustainable Energy................................................................. 316
13.2 Photovoltaic Solar Power ........................................................ 317
13.3 Solar Thermal Power .............................................................. 320
13.4 Hydropower ......................................................................... 321
13.5 OTEC and Thermoclines ........................................................ 324
13.6 Wind Power .......................................................................... 326
13.7 Tidal and Wave Power ............................................................ 327
13.8 Geothermal Power ................................................................. 330
13.9 Biomass Power ...................................................................... 332
13.10 Fusion Power ....................................................................... 333
14 Enhanced End-Use Efficiency .......................................................... 343

14.1 Heat/Cold Storage in Large Buildings ...................................... 344
14.2 Improved Lighting................................................................. 347
14.3 Improved Windows ............................................................... 351
14.4 Heat Pumps .......................................................................... 353
14.5 Improved Appliances ............................................................. 356
14.6 House Doctors....................................................................... 359
14.7 Cogeneration ........................................................................ 364
14.8 Utility Load Management ....................................................... 365
14.9 Energy Storage ..................................................................... 371
15 Transportation............................................................................... 378
15.1 Automobile Energy Basics....................................................... 380
15.2 Corporate Average Fuel Economy (CAFE) ................................. 384
15.3 IC Engines ............................................................................ 386
15.4 Hybrid Cars .......................................................................... 388
15.5 Hydrogen Fuel-Cell Car.......................................................... 390
15.6 Safety................................................................................... 394
15.7 Transportation Potpourri......................................................... 396
16 Energy Economics.......................................................................... 402

16.1 Basic Economics .................................................................... 402
16.2 Discounted Benefits and Payback Periods.................................. 409
16.3 Cost of Conserved Energy ....................................................... 412
16.4 Minimum Life-Cycle Costs ...................................................... 414
16.5 Energy Tax Credits ................................................................. 415
16.6 Petroleum Economy ............................................................... 417
16.7 Imported Oil, Synfuels, and Gasohol......................................... 419
16.8 Plutonium Economy............................................................... 422
P1: OTE/SPH P2: OTE
Contents xi
Appendices
A. Nuclear Arms Chronology............................................................... 431
B. Energy/Environment Chronology ..................................................... 446
C. Units............................................................................................. 454
D. Websites........................................................................................ 460
E. Symbols......................................................................................... 464
F. Glossary......................................................................................... 469
G. Index ............................................................................................ 483

P1: OTE/SPH P2: OTE
Preface
When I arrived on Capitol Hill three decades ago as a science congressional fellow,
I sought a book on the physics of nuclear arms, energy, and the environment. My
bookshelves were full of descriptions of these issues, but not one with the science
behind them. These works could not assist me or any other person having a back-
ground in physics in understanding the “why” of the “scientific process” rather
than the “political process.” So I set out to write such a book. Physics of Societal
Issues is the result.
Decisions guiding policies on nuclear arms, energy, and the environment often
seem mysterious and contradictory, even dangerously wrong-headed. In the mak-
ing of these policies there sometimes appears to be a contest between those who
allow ignorance to prevail and those who are apathetic—between attitudes of “I
don’t know” and “I don’t care.” How else to explain the deployment of MIRVed
ICBMs, a practice that was declared three decades later to be “destabilizing”? The
quest for space-based beam weapons that in reality lack the physical capacity to
carry out their stated mission? The wholesale acceptance of SUVs that use two to
three times the gasoline of an efficient car? And why did it take the oil embargo of
1973–74 to learn that 50 power plants were being wasted to run refrigerators, using
three to four times the necessary energy?
Science-and-technology is the defining force of our age, so it behooves us to
understand the essential processes behind this force that created our society. Most
presidential science advisors have been physicists whose training affords them a
larger view of the physical world. Physicists can play an integral role in creating
solutions to society’s problems, ones that are not prone to errors, either silly or
egregious. Unfortunately, physicists seldom use their skills to examine these issues
and disseminate their conclusions to the public. Physics of Societal Issues is a call to
the broad physics community to join in improving the science-and-public policy
process. It is intended for physicists and engineers who want to understand the
science behind policy issues. Using the techniques and examples offered in this
text, a reader with an adequate knowledge of science at the baccalaureate level
can calculate approximate but useful answers that will inform and enhance the
debate.
xiii
P1: OTE/SPH P2: OTE
xiv Preface
Why This Book?

The elder generation of physicists working on societal issues—Bethe, Drell,
Garwin, Rosenfeld, Panofsky and von Hippel to name a few—have done excellent
work. We will need new physicists of their stature, skill, and dedication to follow
them, but few younger physicists are now participating. In part, this is due to the
fact that physics departments feel they must maintain a high level of competence in
the fundamentals, neglecting the applied issues for another day. The subdivisions
of physics—nuclear physics, particle physics, atomic physics, condensed-matter
physics, astrophysics, biophysics—have many textbooks, while the subdivision
of physics and society lacks even a single text on the three physics-related issues of
arms, energy, and environment. To help remedy this situation, Physics of Societal
Issues is intended for seniors, graduate students, and professionals. This book can
be a starting place for the undergraduate and doctoral students who will work
on applied topics, or for those who just want to know why things are the way
they are.
Fermi’s Envelope
Hans Christian von Baeyer’s book, The Fermi Solution, describes Enrico Fermi’s
elegant approach that simplifies complex physical situations to bare-bones reality.
Physics of Societal Issues applies Fermi’s method to science policy issues. Approx-
imate calculations can be simplistic, but they are the beginnings of a process of
critical thinking toward science policy. Each new situation requires first-principle
thinking to capture its essence. Can nonlinear effects be ignored? Is a spherical ge-
ometry sufficient? Will the parameters valid now remain valid in the future? Has a
“hidden agenda” been used to convert followers? Better algorithms and computers
can improve estimates, but they may not enhance the debate. For example, uncer-
tainties in the main parameters are often more important than calculating higher
order effects. Physics and physicists can provide honest answers that stand up to
the test of time.
Einstein’s Responsibility
“The right to search for truth implies also a duty; one must not conceal any part of what
one has recognized to be true.”
[Albert Einstein, on the Albert Einstein Memorial Statue, Washington, DC]
Einstein is our icon for a no-nonsense, “no-holds-barred” attack on the secrets of

nature, one that is free of personal bias that can fog conclusions. This is the standard
that scientists should follow in both their research and teaching: to be open-minded,
and to refrain from sins of commission and omission. It is tempting to win an
argument by avoiding issues that undercut one’s argument, but scientists must not
P1: OTE/SPH P2: OTE
Preface xv
succumb to the temptation of omission. Our goal in applying the techniques of

physics to issues of science and society must be to examine all the facts and not
conceal any part of what we have recognized to be true.
A no-holds-barred approach to debates on national security can be hindered
by unnecessary government secrecy policies. Such policies prevent advisors with
clearances from publicly using published data sets for fear that a clue will be given
to the actual classified values. This may be reasonable at times, but it can also be
unreasonable if parameters are generally well known or irrelevant. Public debates
can lose substance through blind applications of secrecy rules. The debate on the
Soviet compliance to the Threshold Test Ban Treaty, which affected the debate on
the Comprehensive Test Ban Treaty, is an example of this phenomenon. It was
abundantly clear to me, as the lead State-Department analyst, that the charge of
likely noncompliance was incorrect. The falseness of the noncompliance charge was
easily seen after considering the systematic and statistical errors in the seismic data.
Three Mini-Texts
Part I: National Security. Despite the demise of the Soviet Union, nuclear weapons
continue to be a serious problem as the United States and Russia grapple to
control 2000 metric tons of weapons-usable material and 30,000 warheads. How
many nuclear weapons do the United States and Russia need and what controls
must be imposed on them? Will US nuclear weapons need further testing? If one
country tests, will other countries then begin testing? Can excess plutonium and
weapons-grade uranium be safely stored or destroyed? How much verification
is enough for arms control treaties?
Physics plays a central role in the national security issues discussed in this
text: nuclear weaponry, explosion effects, the neutron bomb, the electromagnetic
pulse, exchange model conflicts, stability of missile-basing, national missile de-
fense, monitoring and verification of arms control treaties, nuclear proliferation,
and terrorism. The national security chapters are as follows:
r Chapter 1. Nuclear Weapons
r Chapter 2. The Offense: Missiles and War Games
r Chapter 3. The Defense: ABM/SDI/BMD/NMD
r Chapter 4. Verification and Arms Control Treaties
r Chapter 5. Nuclear Proliferation
Part II: Environment. Environmental pollution continues to be a major area of con-
cern, one to which physicists can contribute information and analysis. How are
chemical pollutants dispersed in air and water? Can scaling models enhance
our understanding of air quality standards? Can we predict the extent of chem-
ical plumes from fossil power plants and radioactive plumes from accidents
involving nuclear power and nuclear weapons? Can we estimate radon levels
inside buildings or temperatures of radioactive canisters at Yucca Mountain?
Can atmospheric physics determine the fine points on climate change resulting
from CO2 increases? Can medical statistics help or confuse health issues? Are
P1: OTE/SPH P2: OTE
xvi Preface
electromagnetic fields from power lines dangerous? These and other environ-
mental issues are discussed in the following chapters:
r Chapter 6. Air and Water Pollution
r Chapter 7. Nuclear Pollution
r Chapter 8. Climate Change
r Chapter 9. Electromagnetic Fields and Epidemiology
Part III: Energy. The clarion call of the 1973–74 oil embargo appears to have been
forgotten. Will we be able to adapt to increased global petroleum consumption,
rising from today’s 80 million barrels a day (Mbbl/d) to an estimated 110 Mbbl/d
in 2020? Will national security be threatened as it was in the 1991 Gulf War when
we fought to protect oil supplies? The United States now buys as many light
trucks and SUVs as it buys cars, lowering the average mileage of new light ve-
hicles to 24 miles/gallon (mpg). The total light vehicle fleet has an even lower
average of 20 mpg. Can the United States reverse the current trend, which is
moving ever further from the Corporate Average Fuel Economy (CAFE) stan-
dard of 27.5 mpg? Since the oil embargo, the United States has reduced energy
consumption by 50 percent in new cars, new buildings, and new appliances. Is it
technically possible for Americans to reduce consumption by a another 50 per-
cent? Renewable energies are promising, but will they remain just promising?
The 60-percent efficient combined-cycle gas turbine is a bridge to the near future,
but what about natural gas reserves? How much longer before electricity from
solar energy becomes truly economical? The energy chapters are as follows:
r Chapter 10. The Energy Situation
r Chapter 11. Energy in Buildings
r Chapter 12. Solar Buildings
r Chapter 13. Renewable Energy
r Chapter 14. Enhanced End-Use Efficiency
r Chapter 15. Transportation
r Chapter 16. Energy Economics
Physics of Societal Issues is a starting point for these complex issues, and it supplies
references and websites for more in-depth study. The parameters I have used are,
in my opinion, consensus values that can be supported with published results. The
bibliographies at the end of the chapters contain many of these parameters, but not
all. In addition, the first two appendices contain chronologies of events that shaped
the evolution of the issues. History and its past mistakes are often our best guides
to the future.
In writing Physics of Societal Issues, I have drawn on a dozen years of experi-
ence working in government. I have been employed in the US Senate (Committees
on Foreign Relations and Governmental Affairs and as Senator John Glenn’s staff
member for nonproliferation and for the Energy Committee), the State Depart-
ment (Offices of the Under Secretary of State for Arms Control, Nuclear Prolif-
eration Policy, and Strategic Nuclear Policy), the Arms Control and Disarmament
Agency (Strategic Negotiations Division), the National Academy of Sciences (Com-
mittee on International Security and Arms Control), and the national laboratories
P1: OTE/SPH P2: OTE
Preface xvii
at Lawrence-Berkeley (Center for Building Sciences), Argonne (Particle Accelerator

Division) and Los Alamos (Physics Division, and chair of the Director’s Review
Committee for the Nuclear Nonproliferation Division). I gained additional insights
in teaching and research positions at Carnegie Mellon, California Polytechnic, MIT,
Princeton, Stanford, University of Maryland, and the Environmental Protection
Agency. I have gained greatly from my contacts with the American Physical So-
ciety, as chair of both the Panel on Public Affairs and the Forum on Physics and
Society. Perhaps Physics of Societal Issues will encourage readers to move beyond
the classroom to work in these areas. To those who journey on this path, good luck
on your travels.
I have greatly benefited from frequent contact with my physics colleagues in the
American Physical Society’s Forum on Physics and Society, the APS Panel on Public
Affairs, and in government service. I deeply thank those who encouraged me on
this manuscript and its precursor articles in the American Journal of Physics, Scientific
American, and Science and Global Security: Bob Adair, John Ahearn, Barry Berman,
David Bodansky, Aviva Brecher, Bob Budnitz, Ben Cooper, Paul Craig, Alan Crane,
Tony Fainberg, Bob Field, Steve Fetter, Ted Foster, Richard Frankel, Richard Gar-
win, Mike Greene, Ken Haggard, Art Hobson, Allan Hoffman, Ruth Howes, Tina
Kaarsberg, Henry Kelly, Randy Knight, Kevin Kolb, Allan Krass, Barbara Levi,
John Marlier, Matthew McKinzie, John Moulder, Pief Panofsky, John Poling, Paul
Richards, Arthur Rosenfeld, Al Sapperstein, Leo Sartori, Dietrich Schroeer, Peter
Shultz, Steve Smith, Valerie Thomas, Kosta Tsipis, Frank von Hippel, Leonard Wall,
and Peter Zimmerman. In addition, I thank Springer editors, Tom von Foerster,
Jeanine Jordon and Gaurav Jain for their encouragement with this project. I am
greatly indebted to Kathy McKenzie, who devised gentle and eloquent ways to make
the diverse topics of Physics of Societal Issues understandable to those that count, the
next generation of physicists. The cover was elegantly designed by Katherine Bay
to show how Fermi’s back-of-the envelope approach can be used to understand
the non-linear problems of society. I am indebted to Roger Longden for his help
and good humor on the 50–campus tour in 2003. Lastly, without my beloved wife
and companion this book would not have been completed. I am proud to have you
as my friends.
Dave Hafemeister
San Luis Obispo and Washington, DC
P1: GFZ
National Security
1 Nuclear Weapons ............................................................................ 3

1.1 Nuclear Age............................................................................. 3
1.2 Fission Energetics...................................................................... 5
1.3 Scaling Laws and Critical Masses ................................................ 8
1.4 Efficiency and Neutron Generations............................................. 11
1.5 Plutonium Implosion Weapons ................................................... 12
1.6 Boosted Primaries and H Bombs.................................................. 14
1.7 Neutron Bomb.......................................................................... 18
1.8 Nuclear Weapon Effects ............................................................. 20
2 The Offense: Missiles and War Games............................................... 31

2.1 Rocket Equation........................................................................ 31
2.2 ICBM Trajectories...................................................................... 33
2.3 ICBM Accuracy ........................................................................ 34
2.4 GPS Accuracy........................................................................... 36
2.5 Kill Probability = f (CEP, H, Y, R, n, Fratricide) .............................. 38
2.6 Nuclear Conflicts ...................................................................... 41
2.7 Conventional Conflicts............................................................... 50
3 The Defense: ABM/SDI/BMD/NMD ................................................. 55

3.1 ABM History............................................................................ 55
3.2 Target Interactions..................................................................... 57
3.3 Nuclear ABMs.......................................................................... 59
3.4 Particle Beam Weapons .............................................................. 60
3.5 Laser Weapons ......................................................................... 62
3.6 Orbital Chemical Lasers ............................................................. 63
3.7 Earth-Based Lasers .................................................................... 65
3.8 X-ray Laser Pumped with a Nuclear Explosion.............................. 67
3.9 Kinetic Kill Vehicles .................................................................. 70
3.10 Airborne Laser ......................................................................... 72
3.11 AntiSatellite Weapons................................................................ 73
3.12 Rail Guns................................................................................. 74
1
P1: GFZ
4 Verification and Arms Control Treaties............................................... 77

4.1 Verification Context................................................................... 77
4.2 Arms Control Treaties................................................................ 78
4.3 Optical Reconnaissance.............................................................. 81
4.4 Adaptive Optics........................................................................ 83
4.5 Digital Image Processing ............................................................ 84
4.6 Infrared Reconnaissance............................................................. 86
4.7 Radar Reconnaissance................................................................ 88
4.8 Nuclear Tests in Space ............................................................... 90
4.9 Atmospheric Nuclear Tests ......................................................... 92
4.10 Underground Nuclear Tests........................................................ 93
4.11 How Much Verification Is Enough?.............................................. 99
5 Nuclear Proliferation ....................................................................... 105

5.1 Proliferation: Baruch to Iraq........................................................ 105
5.2 Uranium Enrichment................................................................. 116
5.3 Separative Work Units ............................................................... 119
5.4 Nonproliferation in the Former USSR........................................... 121
5.5 Plutonium Production ............................................................... 124
5.6 MTCR and Scuds ...................................................................... 128
2
P1: GFZ
1
Nuclear Weapons
1.1 Nuclear Age

The fission age began in 1932, when James Chadwick discovered neutrons by ob-
serving knock-out proton tracks in a Wilson cloud chamber from the reaction of
alpha particles on beryllium-9,
4
He + 9 Be ⇒ 3 4 He + 1 n. (1.1)
Nuclear weapons are a fluke of nature, since their driving force, a rare isotope
of a rare element, is essentially the only path to producing reactors and nuclear
weapons. Without 235 U one could not produce plutonium without the use of ex-
pensive particle accelerators or other isotopes produced in reactors.
Leo Szilard, the first to consider nuclear bombs, had this recollection on a day in
September 1933:
As I was waiting for the light to change and as the light changed to green and I crossed the
street, it suddenly occurred to me that if we could find an element which is split by neutrons
and which would emit two neutrons when it absorbed one neutron, such an element, if
assembled in sufficiently large mass, could sustain a nuclear chain reaction. I didn’t see at
the moment just how one would go about finding such an element . . . .
Szilard thought that neutron multiplication might take place with beryllium with
the reaction,
1
n + 9 Be ⇒ 2 4 He + 2 1 n. (1.2)
Such a nuclear weapon can’t be built, since Be-fission neutrons lack the necessary
energy to fission 9 Be nucleui. Still, Szilard realized the military importance of chain
reactions even though uranium fission was not discovered for another 6 years.
Since he was greatly concerned that Germany would build and misuse nuclear
weapons, he filed a secret patent of his crude model with the British navy. However,
the sharing of this secret with the navy encouraged its military use, which was
the beginning of Szilard’s contradictory efforts to both militarize and demilitarize
fission. He failed to convince Frederick Joliot not to publish his result that “more
than one neutron must be produced” from uranium fission. After Joliot submitted
3
P1: GFZ
4 1. Nuclear Weapons
an article on this result to Nature on March 8, 1939, Szilard and Fermi published in
the Physical Review that uranium fission produced “about two neutrons.” Thus, in
1939 physicists who read the Physical Review and Nature could deduce that nuclear
weapons could become a reality. Szilard later wrote three letters, all signed by
Albert Einstein, to President Franklin D. Roosevelt: one to warn of Germany, one
to push the Manhattan Project forward, and one to stop the project after Germany
had been defeated. From these modest beginnings, the US stockpile rose from
9 warheads in 1946 to 50 in 1958 to 30,000 in 1965. The Russians followed suit with
40,000 warheads by 1985 (Fig. 1.1). In Szilard’s later years he organized nuclear
scientists into the Council for a Livable World and the Pugwash movement to slow
the nuclear arms race by working peacefully with Soviet scientists.
An excellent history of nuclear weapons and arms control can be found in the
books by Richard Rhodes and Strobe Talbott, listed in the bibliography. Also see
the Nuclear Arms Chronology in Appendix A for a listing of the important his-
torical events. Scientific details of nuclear weapons are classified “top secret” and
“restricted data,” but the basic science of nuclear weapons has been declassified
in Robert Serber’s The Los Alamos Primer, which presents basic equations and con-
cepts, and which was considered worrisome when it became publicized. Lastly,
The Effects of Nuclear Weapons by S. Glasstone and P. Dolan is considered a classic.
Figure 1.1. US and USSR/Russian nuclear stockpile 1945–1996. The Russian curve peaks
in 1985, but the values may be unrealistic since it was driven by public statements from
Minatom Minister Victor Mikhailov. [Natural Resources Defense Council]
P1: GFZ
1.2. Fission Energetics 5
1.1.1 Nuclear Proliferation

At least 24 nations (Section 5.1) attempted to develop nuclear weapons, beginning
with the five nuclear weapon states (NWSs), the “big five” of World War II (the
United States, United Kingdom, Russia, France and China). Nuclear tests by India
in 1974 and 1998, Pakistan in 1998 and North Korea in 2006 increased the list to eight.
South Africa built six uranium weapons, but dismantled them in 1992. During the
Gulf War of 1991, the UN and the IAEA (International Atomic Energy Agency) dis-
covered Iraq’s large nuclear program. North Korea produced enough plutonium
for a few weapons, encouraging South Korea, Japan, and the United States to give
North Korea two commercial reactors in exchange for ending its program and al-
lowing inspections. This 1994 agreement collapsed in 2002 with the announcement
that North Korea had restarted its weapons program in the late 1990s and ejected
IAEA inspectors. Iran moves towards nuclear prowess with its enrichment pro-
gram. In the past, South Korea, Taiwan, Sweden, Switzerland, Brazil, and Argentina
took steps to obtain nuclear weapons, but they stopped. The news is not all dark, as
four states (South Africa, Ukraine, Belarus, and Kazakhstan) have given them up.
The 1970 Non-Proliferation of Nuclear Weapons Treaty (NPT) anchors a global
regime that bans nuclear weapon technologies in non-nuclear weapons states
(NNWSs). The NPT regime relies on declarations by states-parties on their nu-
clear materials, which are monitored by the IAEA inspections to determine their
validity. In return for this loss of sovereignty, NNWSs expect NWSs to greatly
reduce their reliance on nuclear weapons and to negotiate toward their ultimate
elimination. The total elimination of nuclear weapons is unlikely, but substantial
reduction is wise. As part of the bargain, NNWSs believe NWSs must stop test-
ing nuclear weapons to conform to their obligation to stop the nuclear arms race.
The NPT states agreed in 1996 to extend the NPT indefinitely, but only after the
five NWSs stated that they would stop testing and join a Comprehensive Test Ban
Treaty (CTBT). In addition, NWSs are expected to assist NNWSs with their peaceful
nuclear power programs. Beyond these incentives, NNWSs prefer to live next to
non-nuclear neighbors, which has led to the creation of nuclear-weapon-free zones.
The NPT was severely undercut when the US Senate rejected the CTBT by 51 to
48 in 1999 and the Bush administration stated it would not seek CTBT ratification.
The George W. Bush Administration called for the development a 5-kton earth-
penetrating weapon to attack underground bunkers. The congress rejected this
approach, while favoring the reliable replacement warhead.
1.2 Fission Energetics

A 1-Mton weapon can destroy houses at a distance of 5–10 km that give third-
degree burns at 10 km, and lethal radioactive plumes at a distance of 100 km. The
size of 1 million tons of conventional explosive could be likened to a train made
up of cars, each carrying 100 tons of coal. The length of this 1-Mton train would be
(106 tons) (1 car/100 tons) (20 m/car) = 2 × 105 m = 200 km. (1.3)
P1: GFZ
However, the mass of a 1-Mton nuclear weapon is less than one-millionth of this
trainload, making it far easier to deliver nuclear weapons over long distances than
conventional explosives of the same yield.
Fission in thermal (slow) nuclear reactors gives
n + 235 U ⇒ FF1 + FF2 + 2.43n + 207 MeV (1.4)
n + 239 Pu ⇒ FF1 + FF2 + 2.87n + 214 MeV, (1.5)
where FF1 and FF2 are the newly created fission fragments. The total number of nu-
cleons is conserved, which constrains the types of fission fragments produced. The
reactions are the same for fast neutrons in weapons, except that an extra 0.1 neutron
is released. The number of neutrons in a fission event varies between zero and six
with an average energy of 1 MeV, but with some as great as 10 MeV. The fission
fragment range is very short, quickly vaporizing bomb materials. Weapon neutrons
remain very energetic since the explosion takes place in less than a microsecond.
The 207 MeV from the fission of a 235 U nucleus is 20 percent of a nucleon’s mass
energy, going into the following channels:
r fission fragment kinetic energy (168.2 MeV)
r neutron kinetic energy (4.8 MeV)
r prompt gamma rays (7.5 MeV)
r delayed gamma rays and beta rays (14.6 MeV)
r unrecoverable neutrino energy (12 MeV).
The probability of nuclear reactions is given in terms of cross-section, σ , which is
the effective area of a nucleus for a particular reaction. The differential probability
of a neutron experiencing a fissile reaction in a thickness dx is
dP = nf σ dx, (1.6)
where nf is the density of fissile nuclei and σ is given in barns (1 b = 10−24 cm2 ). The
thermal fission cross-section for 235 U is 583 b and for 239 Pu it is 748 b. At weapon
energies of 1 MeV, the 235 U fission cross-section is σ = 1.2 b and for 239 Pu it is σ =
1.7 b. The higher value of σ for 239 Pu at 1 MeV and its extra 0.5 neutron/fission
gives 239 Pu a considerably smaller critical mass than 235 U. Smaller critical masses
are useful for primaries of fusion weapons and for miniaturized multiple warheads
on intercontinental ballestic missiles (ICBMs).
After a 235 U or 239 Pu nucleus captures a neutron, the resultant 236 U or 240 Pu nu-
cleus oscillates like a liquid drop and splits into two fission fragments. The oscilla-
tions are caused by competition between two types of energy: the long-range, repul-
sive, electrostatic interaction between protons and the attractive, short-range, nu-
clear force between neighboring nucleons. The electrostatic force increases with an
increased number of protons in the nucleus, a fact that prohibits stability for nuclei
heavier than lead. The repulsive potential is essentially proportional to the square
of the charge Z, since each proton experiences repulsion from the other (Z − 1)
protons. Since the nuclear radius increases as the cube root of the atomic number
A, the average distance between protons in large nuclei does not vary strongly.
P1: GFZ
1.2. Fission Energetics 7
On the other hand, the short-range, nuclear attractive energy between nucleons
results, to first order, only with its nearest neighbors, which is essentially constant
for nucleons within the nuclear volume. The total nuclear attractive energy is thus
proportional to the number of nucleons, which is approximately proportional to
Z. As nuclei become larger, the repulsive potential grows more than the attractive
potential and these nuclei become unstable. Symbolically we see the Z dependence
below, with the assumption that atomic number A and charge Z are proportional:
Urepulsive α Z2 /r α Z2 /A1/3 α Z2 /Z1/3 α Z5/3 (1.7)
Uattractive α A α Z. (1.8)
Six MeV are released when a neutron is captured by 235 U or 239 Pu, causing os-
cillations and fission. Because target nuclei have an odd number of neutrons, the
binding energy of the absorbed neutron includes pairing energy from combining
spin-up and spin-down neutrons. On the other hand, 238 U with an even number of
neutrons has a smaller neutron binding energy of 4.8 MeV since the pairing energy
is not available. Because less energy is available from neutron capture by 238 U nu-
clei, only fast neutrons over 1 MeV can fission 238 U. Since the 238 U fast-neutron σ is
smaller than it is for 235 U or 239 Pu, 238 U cannot be used to make a fission weapon.
However, energetic 14-MeV neutrons from fusion can fission all isotopes of ura-
nium that can be used in the secondary of a hydrogen bomb. Fission weapons can
also be made with 233 U, which is made in reactors by converting 232 Th to 233 Th,
which then decays to 233 U. The energetic gamma rays from 233 U make weapon’s
production and storage difficult. The IAEA is placing 237 Np and two americium
isotopes under safeguards, since they can in principle be made into weapons.
1.2.1 Fission Energy

The difference in electrostatic energy between protons in new nuclei (236 U or 240 Pu)
and protons in fission fragments accounts for most of fission’s 200 MeV. The repul-
sive self-energy of a uniformly charged 236 U sphere is
U = 0.6(e 2 /4π εo )Z2 /r (1.9)
−15
where e /4πεo = 1.44 MeV-fermi (1 fermi = 1 fm = 10
2
m) and r236 = 1.4A 1/3
fm =
(1.4)(236)1/3 = 8.7 fm, giving
E 236 = 0.6(1.44 MeV-fm)(92)2 /(8.7 fm) = 840 MeV. (1.10)
For simplicity, we assume symmetrical fragments with charge 92/2 = 46, atomic
number 117 [(236–2)/2] and radii 1.4 × 1171/3 = 6.8 fm. The self-energy of the two
fragments is
E FF = 2(0.6)(1.44)(46)2 /6.8 = 540 MeV. (1.11)
The center-to-center distance between the two fragments after fission is twice the
radii, or 2 × 6.8 fm = 13.6 fm, giving a repulsive energy for kinetic energy
E 2FF = (1.44)(46)2 /(13.6) = 224 MeV. (1.12)
P1: GFZ
The change in nuclear-force energy between the initial nucleus and the final two
fragments, E nuc , is obtained from the energy balance,
E 236 = E FF + E 2FF + E nuc (1.13)
840 MeV = 540 MeV + 224 MeV + E nuc .
This gives E nuc = +76 MeV. The sign of E nuc is positive since energy is required to
break nuclear bonds.
The calculated energy E 2FF of 224 MeV for fission fragment repulsion is higher
than its actual value of 170 MeV. This value is reduced because division into fission
fragments is not symmetrical. By using nominal values of Z1 = 56, Z2 = 36, A1 =
142, and A2 = 92, we obtain E 2FF = 200 MeV.
The ratio of fission energy released per nucleus to chemical energy released per
atom is
E nuc /E cbem ≈ (2 × 108 eV/nucleus)/(10 eV/atom) ≈ 2 × 107 . (1.14)
The mass of 235 U nuclei is an order of magnitude larger than combustion masses
CO2 and H2 O, lowering the ratio on a mass basis to E nuc /E chem ≈ 2 × 106 . Since
nuclear weapons are about 20% efficient, the ratio is further reduced to about 106 .
1.3 Scaling Laws and Critical Masses

The largest high-explosive, deliverable bomb was the 10-ton (0.01 kton) bomb that
the United States deployed in the second Gulf War. Nuclear weapon yields have
varied from this level used in backpack weapons for destroying bridges and dams
to the huge Soviet 100-Mton weapon that was tested at the 58-Mton level in 1962. To
help explain the concept of critical masses, we will use scaling laws to consider the
effects of simply changing the size of an object. Scaling laws can answer pragmatic
questions, for example, “Why do cows eat grass and mice eat grains?” To answer
this we may consider animals as simple spherical shapes whose heat loss through
skin is proportional to skin surface area (radius squared). The amount of stored food
energy is approximately proportional to the animal’s volume (radius cubed) and the
type of food it eats and stores. The amount of energy stored divided by the time
between meals is the storage rate, which equals on average the energy loss rate.
This gives an energy balance
Sr 3 = K r 2 , (1.15)
where S and K are the proportionality constants for the storage and loss rates,
respectively. At equilibrium, the radius of the spherical animal is r = K /S. A cow’s
radius is large because it eats energy-poor grass (small S value). If an animal has
thick skin or fur (small conductivity K ), it can be smaller and still gain sufficient
energy from eating grass. On the other hand, mice must eat high quality grains
(large S value) since their area/volume ratio (loss/storage) is large. Scaling shows
that small animals (mice and humming birds) eat often and they must eat energetic
P1: GFZ
1.3. Scaling Laws and Critical Masses 9
grains to overcome their large area/stomach ratio. For the opposite reason, large
animals eat less often and eat less energetic foods, such as grass. Scaling arguments
also show that big animals must have relatively large diameter bones.
1.3.1 Nuclear Scaling Laws

Nuclear fission cross-sections σ determine the likelihood of fission events. Fig. 1.2
below displays fission cross-sections between 0.01 eV and 10 MeV. Note that 235 U
and 239 Pu display significant low energy, thermal cross-sections needed for nuclear
reactors, but that 238 U does not have a thermal cross-section. For this reason, 238 U
cannot be used for reactor fuel. Nor can 238 U be used for fission primaries since
it also lacks an epithermal cross-section below 2 MeV. However, 238 U has σ = 0.6
to 0.9 b (10−24 cm2 ) for neutrons above 2 MeV, which makes the isotope useful
Figure 1.2. Fission cross-section as a function of neutron energy for 235 U, 238 U, and 239 Pu.
The shaded region represents the resonance region where the cross-section varies rapidly.
The boundary of the shaded region does not represent the limit on the peaks and valleys
(Bodansky, 2004).
P1: GFZ
for fast-fission yield in primaries and secondaries once the primary explosion has
begun.
A critical nuclear mass is the minimum mass that will produce a net growth
of neutrons from chain reactions with fast neutrons. The critical mass depends
on isotopic and chemical composition of materials as well as geometry and design
(Fig. 1.3). Following Serber, the critical radius Rc is determined by balancing neutron
loss rate through an area (r 2 ) and neutron production rate in a volume (r 3 ). This
balance exits at the critical radius, obtained from
n[(η − 1)/τ ](4π Rc 3 /3) = (4π Rc 2 )nv, (1.16)
where n is the neutron density, η is the neutron multiplicity per neutron captured in
the fuel, τ is the nuclear-generation lifetime, and v is the average neutron velocity.
This gives the critical radius,
Rc = 3vτ/(η − 1). (1.17)
The lifetime of a nuclear generation is the mean free path λ divided by the neutron
velocity, τ = λ/v. The neutron velocity is obtained from 1 MeV = (1/2)(mc 2 )(v/c)2 ,
giving v = 1.5 × 107 m/s. The value of λ is 1/nf σ , where nf is as before the density
of fissile nuclei and σ is the 235 U fast-fission cross-section of 1.2 b. Values of τ and
λ for uranium at 1 MeV are
τ = λ/v = 1/nf vσ = 1/(5 × 1022 /cm3 )(1.5 × 109 cm/s)(1.2 b) = 10−8 s (1.18)
λ = vτ = (1.5 × 109 cm/s)(10−8 s) = 15 cm. (1.19)
The value of τ = 10−8 s = 10 ns was called a “shake” at Los Alamos during the
war. Substituting these formulas into Rc gives
Rc = 3vτ /(η − 1) = 3λ(η − 1) = 3/nf σ (η − 1). (1.20)
Using τ = 10 ns, v = 1.5 × 109 cm/s, and (η − 1) = 2.5 − 1 = 1.5 gives Rc = 30 cm.
Serber in the Los Alamos Primer obtained 9 cm with a mass of 55 kg, using the
diffusion equation. Reflectors made of beryllium or uranium reduce the critical
mass from the “bare-sphere” values. Implosion increases the density of 239 Pu or
Figure 1.3. Hiroshima

weapon. Principle of 235 U
gun-assembly nuclear device
(Glasstone and Dolan, 1977).
P1: GFZ
1.4. Efficiency and Neutron Generations 11
Table 1.1. Critical mass

SNM Bare sphere U tamper
235 U 47.9 kg 15 kg
239 Pu 10.2 kg 5 kg
235 U and 239 Pu critical masses at normal densities are

given for case of bare spheres and when a natural ura-
nium tamper surrounds the base sphere. [TOPS Task
Force, Nuclear Energy Research Advisory Committee,
Department of Energy, 2000]
233
U, reducing the critical mass (Table 1.1). The proportionality
critical mass = nf Rc 3 α nf /nf 3 α 1/nf 2 (1.21)
shows that the critical mass is inversely proportional to the square of the density
of fissile material nf .
1.4 Efficiency and Neutron Generations

1.4.1 Fission at the Rate of 17 kton/kg
One kilogram of fissile material contains immediate energy
(1/235 kg-mole)(6 × 1026 235 U)(170 MeV/235 U)(1.6 × 10−13 J/MeV)
= 6.9 × 1013 J = 17 kton, (1.22)
with 1 kton = 1012 calories = 4.2 × 1012 J. The Hiroshima gun weapon released
13 kton of energy, consuming 0.8 kg of 235 U. The weapon action was initiated
with a pulse of neutrons from a mixture of 210 Pu. and Be. The 22-kton Nagasaki
implosion weapon consumed 1.3 kg of 239 Pu. Gun-barrel designs are less efficient,
while plutonium implosion weapons can obtain 20% efficiency.
1.4.2 Folding Paper 51 Times

The growth of neutrons in a warhead is analogous to folding paper, as each folding
doubles the thickness. To illustrate, folding a sheet of paper of thickness d 51 times
gives a folded thickness of
D = d(251 ) = (7.6 × 10−5 m)(2.3 × 1015 ) = 170 million km, (1.23)
where d = 0.003 in = 0.076 mm. This is the distance to the sun. To simplify the
calculation, let 210 = 1024 ≈ 103 ; then
D = d(251 ) = d(210 )5 (2) ≈ d(103 )5 (2) = (7.6 × 10−5 m)(2 × 1015 ) = 150 million km.
(1.24)
P1: GFZ
A paper folded eighty times (corresponding to the fission events in a 15-kton

warhead) has a thickness
D = d(280 ) = d(1.2 × 1024 ) = 9.2 × 1016 km = 1000 light years. (1.25)
1.4.3 80 Doublings
The number of 235 U fissions in a 15-kton explosion is
(15 kton)(4.2 J × 1012 /kton)(1 eV/1.6 × 10−19 J)(235 U/1.7 × 108 eV) = 2 × 1024 .
(1.26)
Of the 2.5 neutrons released per fission, 1 maintains steady state production, 0.5 is
lost to space and 1 is available to double the fission rate. The number of neutrons
rises according to N = 2i for i neutron generations. The number of doublings to
fission 2 × 1024 235 U is
i = lnN/ln2 = ln(2 × 1024 )/ln2 = 81 − 1 = 80. (1.27)
One generation was subtracted because the number of neutrons in the last gener-
ation equals the sum of all previous generations. (This corresponds to 55 e-folding
generations.) Working backwards, n = 281 = 2.4 × 1024 .
Following Serber and Mark, the neutron production rate from the diffusion equa-
tion is
dn/dt = −D∇ 2 n + [(η − 1)/τ ]n, (1.28)
where n is density of neutrons, η is neutron multiplicity, and D is the diffusion
coefficient. The first term on the right is the diffusive loss of neutrons to regions
with lower densities and the second term is the neutron production rate. For the
case of uniform neutron density without diffusion, absorption, and edge effects,
the rate of change in n is proportional to n, giving exponential growth,
dn/dt = [(k − 1)/τ ]n, (1.29)
with a solution n = no e (k−1)t/τ . The effective number of neutrons emitted per cap-
tured neutron is k, reducing the multiplicity η to take into account various neutron
losses. A k value of 2 for nuclear weapons gives (k − 1)/τ = 1/10−8 = 108 /s, or a
rise time of 10 ns.
1.5 Plutonium Implosion Weapons

The advance in plutonium weapons has been dramatic. The 1945 “Fat Man” was a
22-kton bomb with a diameter of 1.5 m, while the Peacekeeper’s (W-87) 300-kton
warhead has a diameter of only 0.6 m. Nuclear artillery shells are only 0.16 m in
diameter.
Plutonium is produced when 238 U absorbs a neutron to become 239 U, which
beta-decays in minutes to neptunium (239 Np), which in turn beta-decays in days
P1: GFZ
1.5. Plutonium Implosion Weapons 13
to plutonium (239 Pu):

1
n + 238 U ⇒ 239 U ⇒ 239 Np + e− + ν ⇒ 239 Pu + e− + ν ⇒ 235 U + 4 He, (1.30)
24 min 2.4 day 24,000 yr
where e− is an electron and ν is an anti neutrino. Prior to the atomic age, plutonium
was produced naturally just below the surface of the earth. Strangely, the uranium
at Oklo, Gabon, in Africa, contains only 0.4% 235 U, rather than the usual 0.7%.
This apparent anomaly can be explained because the 235 U content 1.8 billion years
ago was 3% when correcting for 235 U’s half-life of 700 million years, a content
similar to today’s power reactor fuel. The 235 U content was depleted at that time by
its consumption in a natural nuclear reactor, which operated for several hundred
thousand years. The rich uranium deposit was in a damp place with enough water
to moderate neutrons creating a natural reactor without human effort.
Plutonium is favored over highly enriched uranium (HEU with 90% 235 U) for
weapons since it emits more neutrons per fast fission (ν = 2.94 versus 2.53), more
neutrons per neutron capture (η = 2.35 versus 1.93), has a higher fast fission cross-
section (1.7 b versus 1.2 b), and has slightly more energy (214 MeV versus 207 MeV).
For these reasons, plutonium makes smaller primaries, which are essential for mul-
tiple independently targetable reentry vehicles (MIRV) on ICBMs and submarine-
launched ballistic missiles (Section 2.5). However, plutonium is more difficult to
make into nuclear explosives because of the high rate of spontaneously emitted
neutrons that the isotope 240 Pu emits. These precursor neutrons can begin a chain
reaction before plutonium reaches its most compact form. We call this preinitiation,
which is similar to preignition in automobiles, when the spark fires the fuel be-
fore maximum compression. Severe preignition prevents cars from operating; the
technical fix is to delay the spark. In a similar fashion, slow-moving, plutonium
gun-type weapons would preinitiate and lose considerable yield, but this problem
can be overcome by explosive implosion using multipoint detonation. Hollow pits
of plutonium enhance efficiency and allow volume for deuterium–tritium gas to
give a fusion boost. A hydrogen secondary is often attached to the primary and, in
some cases, a dial-yield feature is used to tailor the yield to the mission.
During a long reactor residency, a considerable fraction of 239 Pu captures a sec-
ond neutron to become 240 Pu. The length of stay in a reactor determines whether
the plutonium is weapons-grade Pu (6% 240 Pu, made in a few months) or reactor-
grade Pu (> 20% 240 Pu, made over several years). The isotopic contents of the five
most common types of plutonium are listed in Table 1.2. Plutonium metallurgy is
complicated by the fact that it exists in several different phases, but it is stable in
the delta phase with a small amount of gallium. Note that plutonium from breeder-
reactor blankets is excellent weapons-grade plutonium, while mixed-oxide (MOX)
fuel, used in thermal reactors, is not.
The rate of spontaneous fission neutrons from 239 Pu is 0.022/g · s and from 240 Pu it
is 100,000 times higher at 910/g·s. (The spontaneous fission rate of 235 U is about 1%
of the 239 Pu rate.) The spontaneous neutrons from 5 kg of weapons-grade plutonium
come primarily from 240 Pu, not 239 Pu, as is seen in the calculation
dn/dt = (5000 g)(0.938 × 0.022 + 0.058 × 910)(n/g · s) = 260,000 n/s. (1.31)
P1: GFZ
Table 1.2. Isotopic composition of various grades of plutonium

(% Pu isotope)
Pu grade 238 Pu 239 Pu 240 Pu 241 Pu 242 Pu
Super-grade – 98 2 – –
Weapons-grade 0.01 93.8 5.8 0.4 0.02
Reactor-grade 1.3 60 24 9 5
MOX-grade 2 40 32 18 8
Breeder blanket – 96 4 – –
Mark, 1993.
Neutron multiplication and (α, n) reactions on light impurities (oxygen and ni-
trogen) can marginally increase the neutron rate in metallic plutonium. The sponta-
neous neutron rate is 4 times larger in reactor-grade plutonium (24% 240 Pu), which
itself would be doubled if it were in plutonium oxide form from the (α, n)-reaction.
Clearly it is more difficult to make a warhead with reactor-grade plutonium, but it
can be done.
If a critical mass is assembled slowly in a gun-barrel device, the assembly time
is about
t = d/v = 0.1 m/300 m/s = 0.3 ms. (1.32)
This allows 80 spontaneous neutrons (250,000 n/s × 0.3 ms) during near critical-
ity, considerably lowering the yield. The assembly time of an implosion bomb is
reduced by a factor of 100 compared to the gun-barrel design; this is due to two
things: The critical implosion sizes are less than 10% the length of the gun barrel
and the implosion velocity is more than a factor of 10 higher. The shock velocity in
plutonium is about
vshock = (Ym /ρ)1/2 = (5 × 1011 /20 × 103 )1/2 = 5000 m/s (1.33)
where Ym is Young’s modulus and ρ is mass density. The compression time for an
implosion is less than
t = 0.01 m/5000 m/s = 2 μs. (1.34)
During this time interval about 0.5 neutron is generated (250,000 n/s ×2 × 10−6 s),
and this can be reduced by a factor of three by using super-grade 2% 240 Pu. (See
Section 5.5 for a discussion of proliferation and reactor-grade Pu.)
1.6 Boosted Primaries and H Bombs

1.6.1 Basic Physics
Fusion of hydrogen into helium on the sun sustains planet Earth. Shortly after physi-
cists began forming ideas on fission bombs, they realized that much more explosive
P1: GFZ
1.6. Boosted Primaries and H Bombs 15
energy could be available from fusion, which is the combining of hydrogen isotopes
into helium. The sun fuses four hydrogen isotopes into helium, releasing 27 MeV
in a three-step, proton-burning process:1
1
H + 1 H ⇒ 2 H + e+ + ν (1.35)
1
H + 2 H ⇒ 3 He + γ (1.36)
3
He + 3 He ⇒ 4 He + 21 H. (1.37)
The mass loss in going from four 1 H nuclei (4 × 1.008 amu = 4.032) to one 4 He
(4.003 amu) is 0.029 amu. This converts 0.7% of the original mass to energy, an
amount that is much less than the 100% conversion from antimatter conversion,
but much more than that which results from chemical explosives. The sun’s gravity
confines the energetic hydrogen/helium plasma to high pressures and tempera-
tures. Since the sun has a life span of 10 billion years, it can slowly use a three-step
process for gravitational confinement fusion.
The little time available for plasma machines and nuclear weapons requires them
to use a one-step process to obtain 4 He from deuterium (2 H or D) and tritium (3 H
or T), or “D plus T gives He,” as in
2
H + 3 H ⇒ 4 He + 1 n + 17.6 MeV. (1.38)
The alpha particle carries 3.2 MeV and the neutron carries 14.4 MeV, enough to
fission all the isotopes of U and Pu, as well as 6 Li to produce tritium for further
fusion,
1
n + 6 Li ⇒ 3 H + 4 He. (1.39)
DT fusion develops 5 times more usable energy/mass than does fission, or 5 ×
17 kton/kg = 85 kton/kg, or 12 g/1 kton:
fusion: 17.6 MeV/5 amu ≈ 3.5 MeV/amu (1.40)
fission: 170 MeV/235 amu ≈ 0.7 MeV/amu (1.41)
1.6.2 Fusion Interaction

Nuclear weapons do not have large gravitational or magnetic fields to confine the
hot plasma. But nuclear weapons do have inertial confinement since all the neutron
generations can take place in a microsecond before the weapon blows apart. If
the temperature of a fusion weapon is the temperature of the sun’s interior at
20 million K, the kinetic energy of particles is 2.5 keV and the velocity of a deuteron is
equal to 5 × 105 m/s. In the last 0.1 μs, an unconstrained deuteron travels a distance
d = (5 × 105 m/s)(10−7 s) = 5 cm, but this overstates the distance because scattering
within the plasma shortens the net distance traveled. For fusion to take place, D
and T nucleons must have sufficient kinetic energy to overcome coulomb repulsion
between the D and T nuclei. (See Section 13.10 for discussion of laser fusion at the
1
The carbon–nitrogen fusion process is more likely on the sun, but it gives the same result.
P1: GFZ
National Ignition Facility.) The root mean square radius of the deuteron is 2 fm and
for the triton it is 1.6 fm. However, the wave function extends beyond this by about
0.2–0.4 fm, giving a total separation of about 4 fm. The repulsive potential requires
DT tunneling for the particles to make contact because the repulsive potential is
larger than 5-keV thermal energies at
U = (e2 /4πε o )/(rD + rT ) = 1.44 MeV–fm/4 fm = 360 keV. (1.42)
1.6.3 Tritium Supplies

A boosted primary contains DT gas to magnify its yield. The energy from DT is
small, but the extra neutrons, released early in the cycle, allow the fission cycle
to skip many generations, increasing the fraction of nuclei that fission. Without
tritium, modern nuclear weapons would not function since DT reactions are needed
to raise the yield to ignite the secondary stage.
Since tritium decays (T1/2 = 12.3 yr, mean life τ = 12.3 yr/0.693 = 17.7 yr), it must
be manufactured so that nuclear arsenals are maintained. The United States has
not produced tritium since 1988. After considering proposals to make tritium in
dedicated accelerators or reactors, the Department of Energy (DOE) opted to make
tritium at an existing Tennessee nuclear power plant operated by the Tennessee
Valley Authority (TVA). We estimate tritium demand by considering how much
tritium is required for a given year. The calculation below is for the Strategic Arms
Reduction Treaties (START I–II) and the 2003 Strategic Offense Reduction Treaty
(SORT) and are extended to lower numbers of warheads.
Tritium is produced by the absorption of neutrons by 6 Li in thermal reactors.
DOE stated in 1998 that it would begin tritium production of 2.5 kg/yr by 2005 for
nuclear weapons under START I, which could sustain some 6000 warheads. DOE
stated it would postpone production to 2011 at 1.5 kg/yr if START II entered-into-
force with a limit of 3500. Deeper cuts in warheads relax tritium requirements and
would further postpone the need for new tritium. Here we estimate the production
time-delay for a wide variety of arms control treaties, but these estimates ignore
the details of the actual tritium cycle (reserves, pipeline, recycle losses, decay in the
warhead). In 2005, under START I, the tritium that will be needed in the stockpile
under steady-state conditions is
m1 = τ (dm/dt) = (17.7 yr)(2.5 kg/yr) = 44 kg. (1.43)
In 2011, under the now discarded START II, the tritium needed would be
m2 = τ (dm/dt) = (17.7 yr)(1.5 kg/yr) = 27 kg. (1.44)
For START I, the average amount of tritium per warhead is, perhaps,
44 kg/10,000 warheads = 4.4 g/warhead, (1.45)
The energy released is small at one-third a kilo ton (4.4 g × 1 kton/12 g). But a pulse
of a mole of neutrons (6 × 1023 ) rapidly advances the number of neutron genera-
tions, increasing fission yield and lowering the requirement for fissile material. The
years in which additional tritium would be needed under SORT and other nuclear
P1: GFZ
1.6. Boosted Primaries and H Bombs 17
force regimes is estimated below:
SORT (START III) at 2000 warheads + 2000 reserves:

m3 = (4000 warhead)(5 g/warhead) = 20 kg (1.46)
t = −ln(m3 /m1 )(τ ) = −[ln(20 kg/44 kg)](17.7) = 14 yr; 14 + 2005 = 2019. (1.47)
SORT at 2000 warheads: m4 = 10 kg, t = 26 yr; 26 + 2005 = 2031. (1.48)
SORT II (START IV) at 1000 warheads: m5 = 5 kg, t = 38 yr; 38 + 2005 = 2043.
(1.49)
SORT III at 200 warheads, when UK/China/France are part of the regime:
m6 = 1 kg, t = 67 yr; 67 + 2005 = 2072. (1.50)
1.6.4 Radiation Compression

From the beginning of the Manhattan Project, Edward Teller wanted to develop the
hydrogen bomb, so he refused to work on pure fission bombs. The initial idea was
to compress the secondary with mechanical shock waves, but the primary destroys
the secondary before sufficient compression takes place. However, Stanislaus Ulam
and Teller deduced that radiation pressure from the primary’s x-rays could com-
press and heat the secondary to ignite fusion. In the late 1970s, Howard Moreland
published rough drawings of the hydrogen bomb. The government’s case to pre-
vent publication was greatly weakened when it was discovered that it had al-
ready declassified these facts, and they were publicly available at the Los Alamos
library.
A primary that generates 10 kton in 100 ns has a thermal power
P = (10 kton)(4 × 1012 J/kton)/(10−7 s) = 4 × 1020 W, (1.51)
which gives a radiant flux from a 10-cm radius primary of
P/Area = 4 × 1020 W/0.01 m2 = 4 × 1022 W/m2 = εσ T 4 . (1.52)
This corresponds to a black-body temperature of some 20 million K, similar to

the sun’s interior temperature. The x-ray distribution (normalized to sun’s 6000-K
surface at λ = 0.5 μ) peaks at
λ = (6000 K/20 M K)(0.5 μ) = 0.15 nm, (1.53)
which gives kilovolt x-rays (hc/λ). The x-rays are absorbed by the secondary to
reradiate new x-rays, which vaporize outward and implode inward from the abla-
tive shock.2
2
Since the x-rays penetrate only a small distance, they vaporize the small volume near the
surface. As the vaporized material can only go away from the surface, it gives an ablative
shock to the system.
P1: GFZ
1.6.5 Lithium Deuteride

The first hydrogen bomb, “Big Mike,” was a large thermos bottle containing liquid
deuterium, giving a yield of 10 Mton. Mike could not be delivered with bombers
or ICBMs because of the extreme size of its cryogenics. A deliverable H bomb was
soon developed with deuterium in the LiH salt in the form of 6 Li2 H. Since this salt
is a solid, there is no need for a cryostat. The neutrons interacting with 6 Li gave
instant tritium to interact with deuterium.
1.7 Neutron Bomb

The neutron bomb produces smaller blasts to reduce collateral damage. At the same
time, neutron bombs enhance neutron emissions to more effectively kill troops,
producing prompt deaths in 5 min with a fast neutron dose of 80 Sv (1 Sv =
100 rem, 80 Sv = 8000 rem). The introduction of the neutron bomb was a shift from
pure fission tactical weapons to weapons that were approximately equally divided
between fission and fusion at 1 kton. Fusion produces more neutrons per energy
released and these neutrons are much more energetic at 14 MeV as compared to
fission’s 1–2 MeV neutrons.
The political debate on deploying neutron bombs was sharply contested in the
United States and Europe in the late 1970s. Those who wanted to deploy neutron
bombs were concerned that tactical weapons would not be used in Europe because
their yields would be deemed too damaging, particularly in Germany. They wanted
1-kton weapons that would incapacitate tank crews at a distance of 850 m, compared
to 375 m for a pure fission 1-kton weapon. Proponents believed that deployment of
the neutron bomb would increase Soviet perceptions that the United States would
actually use it, a result that would deter a Soviet invasion. On the other hand,
it could be envisioned that the deployment of neutron bombs would lower the
psychological and bureaucratic threshold for first-use of nuclear weapons. Such a
deployment would increase the probability of its first-use by local commanders,
thus starting a more general nuclear war. Lastly, there already was considerable
deterrence to discourage an invasion because the United States had other nuclear
weapons in Europe. The view of the opponents carried the day as Congress blocked
its deployment. Those who lost the debate commented that a different name, the
reduced-blast bomb, would have helped their case.
1.7.1 Neutron Effluence

A 1-kton weapon has yield energy
Y = (4.2 × 1012 J/kton)(1 eV/1.6 × 10−19 J) = 2.6 × 1031 eV. (1.54)
If one of the three neutrons from a fission event were to escape the warhead, the
number of neutrons exiting a 1-kton fission weapon would be
Nfission = (2.6 × 1031 eV)(1 n/2 × 108 eV/fission) = 1.3 × 1023 n. (1.55)
P1: GFZ
1.7. Neutron Bomb 19
The 14.4 MeV fusion neutrons predominately exit the weapon, since they are very
energetic and the bomb case is designed for this purpose. The number of neutrons
exiting a 1-kton pure fusion weapon is
Nfusion = (2.6 × 1031 eV)(1 n/1.76 × 107 eV/fusion) = 1.5 × 1024 n. (1.56)
The value for fusion is a factor of 10 greater than that for pure fission. If a 1-kton
neutron bomb consists of 50% fission to initiate the remaining 50% fusion, the
number of exiting neutrons is
NNBomb = (1.5 × 1024 + 1.3 × 1023 )/2 = 8.2 × 1023 n, (1.57)
which is 6 times that of a pure fission weapon. Neglecting scattering and absorption
by the atmosphere, the neutron fluence from a pure fission bomb at 800 m is
f l fission = (1.3 × 1023 n)/(4π )(8 × 104 cm)2 = 1.6 × 1012 n/cm2 (1.58)
and 1013 n/cm2 for the neutron bomb.
1.7.2 Radiation Dose

Neutron bombs produce only one-half as much radioactivity as pure fission bombs,
but they give a much larger radiation dose to close-in troops. A 1-kton tactical fission
weapon delivers a neutron dose of 2.5 Sv (250 rem) at a distance of 800 m, which
takes into account neutron scattering by air. The radiation dose from a neutron
bomb is considerably greater because it produces 6 times more neutrons, and fusion
neutrons are 7 times more energetic than fission neutrons (14 MeV/2 MeV). Neutron
bombs deliver a dose at 800 m of
(2.5 Sv/kton-fission)(6 flux)(7 energy) = 100 Sv = 104 rem. (1.59)
1.7.3 Reduced Blast

Most fission energy appears as fission-fragment kinetic energy, which heats bomb
debris to produce a blast wave. Pure fusion contributes less blast energy since
escaping neutrons carry considerable energy away from the weapon. Reduced
damage to buildings from neutron bombs is due to the fact that blast energy is
50% of pure-fission yield, while it is 20% of pure-fusion yield. Thus, the blast
energy for a 1-kton neutron bomb is (0.5 + 0.2)/2 = 0.35 kton-blast/kton. A house
is destroyed with a threshold overpressure ( po ) of 5 psi. Blast overpressure ( p) falls with
the third power of distance, or p = BY/r 3 (Section 1.8), where Y is yield and B is a
constant. This gives a survivability radius rs which is proportional to the cube root
of the yield, or rs = (BY/ po )1/3 . Thus, the ratio of blast destruction areas (neutron
bomb/fission bomb) is
blast: ANBomb /Afission = (0.35 kton/1 kton)2/3 = 0.5, (1.60)
showing that destruction area is reduced by 50% by using neutron bombs. If the
motive is to kill tank drivers but let buildings survive, the relative effectiveness of
P1: GFZ
the two weapons is determined by dividing the ratio of neutron-kill area by the
ratio of blast area for each weapon type:
neutron bomb/fission bomb = (n dose: ANBomb /Afission )/(blast: ANBomb /Afission )
= (800 m/375 m)2 /0.5 = 4.6/0.5 = 9. (1.61)
Thus, the neutron bomb was 9 times (per unit area destroyed) more effective at
killing tank drivers as compared to tactical nuclear weapons. However, congress
blocked the neutron bomb since it was more concerned that its deployment would
lower the political threshold for the first use of nuclear weapons, which could initiate
general nuclear war.
1.8 Nuclear Weapon Effects

Nuclear weapon energy appears as blast pressure waves, thermal radiation, and
prompt/delayed radiation. The division of the total energy into these quantities
depends on weapon yield, ratio of fusion to fission energies, and height of burst.
This chapter discusses blast, thermal, and radiation effects, as well nuclear win-
ter and electromagnetic pulses, while Chapter 7 discusses low-dose radiation ef-
fects. Nuclear weapons can destroy opponent’s weapons, but they more easily
can devastate cities and people. The 15–20 kton weapons used on Hiroshima
and Nagasaki killed 180,000 people, about 40% of the inhabitants. Those that
died from radiation also died because they were within the lethal blast radius.
Outside the lethal blast areas about 400 Japanese died from delayed, low-dose
cancer. However, the 15-Mton Castle Bravo hydrogen bomb created a gigantic
radioactive plume far beyond the blast, killing Japanese fisherman on their ves-
sel, the Lucky Dragon. The destructive effects of Mton-size weapons could be
immense under (hopefully rare) United States and Russia forces that operate
under launch-on-warning scenarios. In the 1960s, Secretary of Defense Robert
McNamara defined mutual assured destruction as the assured second strike that
would kill 25% of a nation’s population and 50% of its industry, as shown in
Table 1.3.
Table 1.3. Assured destruction (number of hard-target

warheads to kill 25% of population and destroy 50%
of industry)
1999 population W88 warheads
China 1281 M 368
Iran 64 M 10
Iraq 21 M 4
North Korea 22 M 4
Russia 152 M 51
United States 259 M 124
McKinzie, 2001.
P1: GFZ
1.8. Nuclear Weapon Effects 21
Table 1.4. LD-50 radii (distances from the explosion where 50% of
the affected population would die from blast, thermal radiation, and
nuclear radiation from 1-Mton surface and air explosions)
1 Mton Blast (5 psi) Thermal (7 cal/cm2 ) Radiation (4.5 Sv, 450 rem)
Surface 4.6 km 11 km 2.7 km
Air burst 6.7 km 17 km 2.7 km
Glasstone and Dolan, 1977.
Some radiation rules-of-thumb effects are as follows:

r lethality from neutrons predominates up to a few kilo ton;
r lethality from blast pressure waves predominates from 5–100 kton;
r lethality from thermal radiation predominates above 100 kton, but lethal radiation
plumes can extend considerably beyond 100 miles.
Fusion makes much less radioactivity than fission since it does not produce
fission fragments. Fusion neutrons are very harmful to people within 1 km of the
blast, but this effect is much less significant than close-in blast effects and fission
radioactivity. The yield of a secondary stage is about 50% fusion and 50% uranium-
fission. Thus, a 1-Mton, 50–50 weapon has about 500 kton of fission, while a 10-kton
primary is 100% fission. Blast height is extremely important in determining the
amount of radioactive fallout. If an explosion takes place at low altitudes, excess
neutrons produce large amounts of radioactivity in the soil, which disperses in
a plume. High altitude bursts make much less radioactivity since nitrogen and
oxygen absorb the neutrons, which decay quickly, but the 14 C lingers. In addition,
a high altitude burst directly disperses and dilutes the radioactivity. See Table 1.4.
A particularly nasty target would be a nuclear reactor. One estimate for a 1-Mton
bomb hitting a 1-GWe reactor predicts an area of 34,000 km2 , which would give a
lifetime dose of over 1 Sv (100 rem) to the affected population.
1.8.1 Overpressure
Overpressure of 5 psi (30 kPa) destroys wood and brick houses beyond repair.
One might think that blast pressure would diminish as the inverse square of the
distance since a blast energy pulse is approximately the peak overpressure ( p) times
the volume element (V):
Y = ApV = Ap(4πr 2 r ), (1.62)
where A is a constant and Y is weapon yield. If the pressure pulse width, r ,

were constant, the pressure would, to first order, fall as r −2 , but super shocks
enhance nonlinear processes, directing considerable energy into higher harmonics.
The different wave velocities for the harmonics cause dispersive broadening of
the pulse-width in proportion to radial distance traveled (r α r ). Inserting r
P1: GFZ
Figure 1.4. High overpressure. Peak overpressure on the ground for a 1-kton burst (high
pressure range) (Glasstone and Dolan, 1977).
proportional to r into Eq. 1.62 gives a pressure of
p = C(Y/r 3 ) (1.63)
where C is a constant depending on units.

A numerical fit to the empirical data in Glasstone and Dolan gives
p = 14.7(Y/r 3 ) + 12.8(Y/r 3 )1/2 in psi, megaton, nautical miles (1860 m) (1.64)

p = 6.31(Y/r ) + 2.20(Y/r )
3 3 1/2
in atmospheres, megaton, kilometers. (1.65)
The first term is sufficiently accurate for small distances when attacking silos
(Fig. 1.4), but it is not accurate for the greater expanses of cities (Fig. 1.5). See
Table 1.5.
The maximum distance at which an explosion will just destroy an object is de-
termined from the object’s critical pressure of destruction, called hardness (H).
The critical distance (rc ) in the high-pressure region is obtained from setting
P1: GFZ
Figure 1.5. Low overpressure. Peak overpressure on the ground for a 1-kton burst (low
pressure range) (Glasstone and Dolan, 1977).
H = CY/rc 3 , giving
rc = (14.7Y/H)1/3 in nautical miles, megaton, psi (1.66)
rc = (6.3Y/H)1/3 in kilometers, megaton, atmospheres. (1.67)
Table 1.5. Overpressure (using one and two

terms for a 1-Mton explosion at distances of 0.1
and 1 nautical miles (1860 m))
Table 186 m 1.86 km
One term 14,700 psi 14.7 psi
Two terms 15,105 psi 27.5 psi
1.8.2 Fallout
The radioactive plume from a nuclear weapon depends on yield, height of blast, and
wind conditions. A 1-Mton weapon can produce a plume that deposits radiation at
a time-integrated level of 5 Sv (500 rem) over an area 30 miles wide and 1000 miles
long. A prompt dose of 4.5 Sv (450 rem) is lethal to about 50%, and essentially no one
P1: GFZ
Figure 1.6. Stratospheric inventory of 90 Sr (Glasstone and Dolan, 1977).
survives 1000 rem. If citizens stay inside buildings, the dose is reduced by a factor
of three. Terrorist dirty-bombs are discussed in Chapter 7, along with calculations
of radioactive plumes from nuclear accidents. See Figs. 1.6–1.9 for data on plumes
and Fig. 1.10 for electromagnet pulse effects.
Figure 1.7. Bikini fallout. Total accumulated dose contours in rads, 4 days after the BRAVO
test explosion (Glasstone and Dolan, 1977).
P1: GFZ
Figure 1.8. Detroit fallout. Main fallout pattern after a 1-Mton surface explosion in Detroit,
with a uniform, steady 15-mile/h wind from the Northwest. The 7-day accumulated dose
contours (without shielding) are for 3000, 900, 300, and 90 rem. The constant wind would
give lethal fallout in Cleveland and 100 rem in Pittsburgh (Office of Technology Assessment,
1979).
P1: GFZ
Figure 1.9. Attack on SS-19s at Kozelsk. Under this scenario 13 million die from radiation
received in the first two days (McKinzie, 2001).
Figure 1.10. Electromagnetic pulse. Schematic representation of an electromagnetic pulse

(EMP) from a high-altitude burst. Fission fragments release prompt megaelectronvolt
gamma rays, which interact with the thin upper atmosphere creating Compton electrons.
The electrons spiral with reasonable coherence in the Earth’s magnetic field at about 1 MHz
because they begin at essentially the same time. The October 1962 test at Johnston Island
shut down the power grid in Hawaii and blocked radio and TV for several hours on the
West Coast and throughout the Pacific region (Glasstone and Dolan, 1977). (See problems
1.17 and 1.18.)
P1: GFZ
Problems 27
1.8.3 Cratering
The radius of a crater in hard rock is about r = 160 Y0.3 , where r is in meters and Y
is in megatons. This gives radii of 150 m for 1 Mton and 75 m for 100 kton (OTA).
1.8.4 Nuclear Winter

The volcanic eruption of April 10, 1815, on Tambora led to global cooling and June
frosts in 1816, the “year without a summer.” Atmospheric physicists realized in
1982 that large nuclear attacks on cities create massive amounts of micron-sized
soot and raise them to the stratosphere, with effects similar to very large volcanic
eruptions. It was projected that 10,000 0.5-Mton warheads could reduce light levels
to a few percent of ambient levels and temperatures could drop by 30◦ C for a month,
warming to 0◦ C for another 2 months. Hence, the name nuclear winter. A key factor is
the lofting of soot to the upper troposphere. Weapons over 0.3 Mton raise soot high
enough to absorb sunlight and heat the upper atmosphere by 80◦ C, which raises
(lofts) the soot higher. A major effect of such weapon blasts would be the destruc-
tion of much of the world’s food supply by low temperatures. The US government
carried out burning and chemical explosions to test some of these ideas, but it is dif-
ficult to test these scenarios peacefully. The debate became one of a matter of degrees
between a nuclear winter and a nuclear autumn. See Ch. 8 for atmospheric models.
Problems
1.1 NPT. (a) What are the trade-offs for the NWSs and NNWSs as members of
the Nuclear Non-Proliferation Treaty? (b) Which events since 1990 have been
positive and which have been negative for the stability of the NPT regime?
(c) What can the IAEA monitor and what are the limitations?
1.2 HEU versus Pu. (a) Which isotope is harder to obtain and why? (b) Which is
easier to make into a weapon? (c) Which is easier to dispose of?
1.3 Neutron rich. (a) Why are heavy nuclei neutron-rich? (b) Why does this create
radioactive fission fragments?
1.4 Neutron moderators. (a) How many head-on collisions must a 1-MeV neutron
have with 1 H, 2 H, and 12 C to become thermalized? (b) Why is 2 H preferable
over 1 H as a moderator?
1.5 0.1n/MeV. (a) An additional 0.1 neutron is released for each 1 MeV of neutron
energy. Is this consistent with the binding energy of the last neutron? (b) How
many fission neutrons are released after the capture of a 14.4 MeV neutron
and does this number effect the secondary yield?
1.6 Asymmetric fission. What is the kinetic energy obtained from a 236 U that splits
into 3/1 mass–ratio fragments?
1.7 Doubling and e-folding. (a) How many doublings and e-foldings does a
1-kton primary need? (b) Why does not this approach work for 500-kton
weapons?
P1: GFZ
1.8 Reactor-grade Pu. (a) What is the neutron emission rate from 5 kg of reactor-
grade plutonium? (b) How many neutrons are emitted during a gun-type
assembly? (c) How many are emitted during an implosion event?
1.9 Fusion neutrons. Determine the kinetic energies of the neutron and alpha
particle after DT fusion.
1.10 Mean free path. (a) What is the mean free path of thermal neutrons and fast
neutrons in delta-phase plutonium with density 15 g/cm3 ? (b) What is the
mean free path and the bare-sphere critical mass if density is doubled?
1.11 Reradiation. How much absorbed energy per square centimeter within
0.05 mm of an Al surface would it take to reradiate at 10 M K?
1.12 H bomb. (a) Why is lithium-6 deuteride (6 Li2 H) useful for hydrogen weapons?
(b) How was the first hydrogen device, Mike, made without LiD or tritium?
(c) Assume a 500-kton weapon gets its energy 50% from fission and 50% from
fusion. How much 235 U is destroyed? How much LiD is destroyed? (d) What
is the volume of the secondary if efficiency is 30%? The density of uranium is
20 and LiD is 0.9.
1.13 Tritium. (a) How much energy does tritium contribute in a primary if 1–10 g are
fused to deuterium. (b) How many neutrons does tritium contribute? (c) What
is the ratio of energies and neutrons with and without tritium? (d) How many
generations do DT neutrons contribute if they are inserted at the beginning of
the reaction?
1.14 Neutron bomb. Redo the calculation in the text, but let the fission/fusion ratio
be 1/3.
1.15 Pressure at a distance. (a) What is the pressure at 1, 5, and 10 km from a
500-kton explosion. (b) What is the minimum survival distance for houses of
hardness 5 psi?
1.16 Meteor Crater. My DC neighbor Phil Barringer owns Meteor Crater, which
was created 50,000 years ago by an iron–nickel meteor of 50-m diameter and
20-Mton energy, making a crater of 1200-m diameter and 200-m deep. (a) As-
suming the meteor had a density of 3 and a velocity the same as Earth’s or-
bital velocity, what is meteor’s energy? Do energy and crater radius scale with
the cratering formula? (b) What maximal height could 20 Mton raise crater
ejected material? Is it possible that Mars ejecta landed on Earth? (c) What
was the mass of the 100-million Mton meteor that landed at Yucatan and de-
stroyed the dinosaurs 65 million years ago? (d) Asteroid 1950A with a 1-km
diameter is predicted to have a 0.3% chance of hitting the Earth on March
16, 2880. Use generalized equations to describe several ways to deflect the
asteroid? (e) The dinosaurs were annihilated 65 million years ago from a me-
teor that produced a 200-km diameter crater in Mexico. What magnitude ex-
plosion has the energy (megaton) and mass (kilograms at 10 km/s) of the
meteor?
1.17 Electromagnetic Pulse range, Signal-generated EMP. (a) At what altitude
should an EMP take place to cover 50% and 100% of the lower 48 states? (b)
Describe how a blast of x-rays can give a signal-generated EMP on a satellite
or reentry vehicle?
P1: GFZ
Bibliography 29
1.18 EMP fancy. Determine the EMP frequency spectrum and energy fluence from
high-altitude nuclear explosions (Hafemeister, 1983).
1.19 Nuclear winter. (a) Show that transmitted solar flux is diminished by soot
absorption, s = so e −α/ cos θ , here the optical depth is α and θ is the angle be-
tween Sun and zenith. (b) Determine transmitted sunlight using parameters
developed by M. MacCracken in 1988: The sun is at θ = 60◦ , 2000 weapons of
0.5 Mton burn area of 300 km2 /Mton, which is reduced by a factor of 3 from
overlap; 30 kg/m2 flammable material; 2% soot production; 1% of soot arises
to the troposphere; α = (area soot density)(soot absorption of 10 m2 /g).
Bibliography
Beschloss, M., and S. Talbott (1993). At the Highest Levels, Little Brown, Boston, MA.
Bodansky, D. (2004, 2nd edition). Nuclear Energy: Principles, Practices and Prospects, AIP Press,
New York.
Brode, H. (1968). Review of nuclear weapon effects, Ann. Rev. Nucl. Sci. 18, 153–202.
Cochran, T., W. Arkin, M. Hoenig, R. Norris, and J. Sands (1984–1989). Nuclear Weapons
Databooks, Ballanger, Cambridge, MA.
Cochran, T., R. Norris, and O. Bukharin (1995). Making the Russian Bomb, Westview, Boulder,
CO.
Dowling, J. and E. Harrel (Eds.) (1986). Civil Defense: A Choice of Disasters, American Institute
of Physics Press, New York.
Duderstadt, J. and F. Moses (1982). Intertial Confinement Fusion, Wiley, New York.
Glasstone, S. and P. Dolan (1977). The Effects of Nuclear Weapons, Departments of Defense/
Energy, Washington, DC.
Hafemeister, D. (1983). The arms race revisited: science and society test VIII. Am. J. Phys. 51,
215–225.
Hafemeister, D. (Ed.) (1991). Physics and Nuclear Arms Today, American Institute of Physics
Press, New York.
Harwell, M. (Ed.) (1984). Nuclear Winter, Springer-Verlag, New York.
Hewlett, R. and O. Anderson (1966). The New World: 1936–46, US Atomic Energy Commis-
sion, Washington, DC.
Hewlett, R. and F. Duncan (1969). Atomic Shield: 1947–52, Penn State University Press, College
Park, PA.
Hewlett, R. and J. Holl (1989). Atoms for Peace and War: 1953–61, University of California
Press, Berkeley, CA.
Holloway, D. (1994). Stalin and the Bomb, Yale University Press, New Haven, CT.
Kaplan, F. (1978). Enhanced-radiation weapons, Sci. Am. 238(5), 44–51.
Lourie, R. (2002). Sakharov, Brandeis Univ. Press and Univ. Press of New England, Hanover,
NH.
National Research Council (1989). The Nuclear Weapons Complex, National Academy Press,
Washington, DC.
MacCracken, J. (1988). The environmental effects of nuclear war, in Nuclear Arms Technologies
in the 1990s, D. Schroeer and D. Hafemeister (Eds.), American Institute of Physics Press,
New York, pp. 1–18.
Mark, J. (1993). Explosive properties of reactor-grade plutonium, Sci. Global Secur. 4, 111–128.
P1: GFZ
McKinzie, M., T. Cochran, R. Norris, and W. Arkin (2001). The US Nuclear War Plan: A Time
to Change, Natural Resources Defense Council, Washington, DC.
Office of Technology Assessment (1979). The Effects of Nuclear War, OTA, Washington, DC.
Rhodes, R. (1995). Dark Sun: The Making of the Hydrogen Bomb, Simon and Schuster, New
York.
——— (1988). The Making of the Atom Bomb, Simon and Schuster, New York.
Schroeer, D. and J. Dowling (1982). Resource letter: Physics and the nuclear arms race, Am.
J. Phy. 50, 786–795.
Schroeer, D. and D. Hafemeister (Eds.) (1988). Nuclear Arms Technologies in the 1990s, Amer-
ican Institute of Physics Press, New York.
Serber, R. (1992). The Los Alamos Primer, University of California Press, Berkeley, CA.
Talbott, S. (1979). Endgame: The Inside Story of SALT-II, Harper, New York.
——— (1984). Deadly Gambits, Knopf, New York
——— (1988). The Master of the Game, Knopf, New York.
Taylor, T. (1987). Third-generation nuclear weapons, Sci. Am. 256(4), 30–38.
Turco, R., O. Toor, T. Ackerman, J. Pollack, and C. Sagan (1983). Nuclear winter: Global
consequences of multiple nuclear explosions, Science 222, 1283–1292.
——— (1990). Climate and smoke: An appraisal of nuclear winter, Science 247, 166–176.
von Hippel, F. and R. Sagdeev (Eds.) (1990). Reversing the Arms Race, Gordon and Breach,
New York, 1990.
P1: GFZ
2
The Offense: Missiles and War Games
2.1 Rocket Equation

The first German V2 flight of October 2, 1942, was the first of 3700 more V2
flights. The V2 carried 750 kg a distance of 300 km, similar to today’s Russian Scud B.
The accuracy of the V2 was poor, only 35% landed within 2 km of their targets. At
this rate, the accuracy of ICBMs would be 60 km over a range of 10,000 km. As
ICBMs improved, the nuclear arms race shifted from production of slow, recallable
bombers to that of fast, nonrecallable, MIRVed (multiple, independently targetable
reentry vehicles) ICBMs. The increased accuracy of ICBMs led to decreased weapon
yields, dropping from multi megatons to about one-half a megaton (Mton). To en-
hance attacks on leadership the United States also developed earth-penetrating
warheads.
If no external forces (gravity or drag)1 bear on a rocket, the total momentum of
the gas and missile is conserved, giving
Fexternal = 0 = d P/dt = m(dv/dt) + Vex (dm/dt), (2.1)
where v is the velocity of the missile, Vex is the exhaust velocity of the gas with
respect to the rocket, and dm/dt is the propellant-mass exhaust rate. The first term
on the right, m(dv/dt), is “ma” for accelerating the rocket mass and remaining fuel,
and the second term is the “momentum-thrust” term. This gives dv = −Vex (dm/m),
which integrates to
vf = vo + Vex ln(mo /m f ), (2.2)
where vo is the initial velocity, vf is the final velocity, mo is the initial launch weight,
and mf is the final (throw) weight. The logarithmic term reduces the effectiveness
of extra propellants, but this can be partially overcome by using several stages to
reduce the mass of each succeeding stage. Exhaust velocity depends on the fuel type
1
Drag force = 0.5ρv2 ACd , where ρ is air density and A is cross-sectional area. The drag
coefficient Cd is less than 0.5 at subsonic speeds, but increases sharply by a factor of 2–3
above the speed of sound. The lift force has the same appearance, except the lift coefficient
is smaller for missiles.
31
P1: GFZ
32 2. The Offense: Missiles and War Games
and the nozzle configuration. Liquid fuels are faster at 3.6 km/sec, but solid-fueled
rockets, in spite of their smaller 2.7-km/s exhaust velocity, are preferable for their
quick response, longevity, safety, and reduced maintenance. The former Soviets
had a difficult time perfecting solid-fueled rockets, as they continued to use liquid
fuels for SS-18s and submarine-launched ballistic missiles (SLBMs). Solid-fueled
rockets are difficult to control since the inability to close valves on the fuel requires
that all the fuel be used. Solid-fueled missiles make complicated maneuvers to use
excess fuel, but still maintain excellent accuracy. Another approach is to explosively
blow out the missile’s sidewall to release gas at the proper moment. “Specific
impulse” is used rather than exhaust velocity as a parameter and it is defined as
Isp = Vex /g where g is the acceleration of gravity. The American Physical Society’s
Directed Energy Weapons Study used Isp = 306 s for the liquid-fueled SS-18, giving
Vex = Isp g = (306 s)(9.8 m/s) = 3.0 km/s.
The theoretical launch-weight to throw-weight ratio for a one-stage strategic
rocket with a velocity of 7 km/s is
mo /mf = exp[(vf − vo )/Vex ] = exp[(7 km/s)/3 km/s)] = 10. (2.3)
In practice this ratio is 20–30 because of the inefficiencies of wasted propulsion,
air drag and gravity. For a vertical launch in vacuum above a flat Earth, the final
velocity is
vf = vo + Vex ln(mo /mf ) − gt burn , (2.4)
where fuel burn-time is tburn . The burn-time ranges from 1 min for a fast-burn
booster to 5 min for an SS-18.
2.1.1 V2
The one-stage V2 had a terminal velocity 1.6 km/s, which delivered its payload at
a range of
X = vf2 /g = (1600 m/s)2 /(10 m/s2 ) = 300 km. (2.5)
The V2 had a launch-weight of 12.8 tonn (1 tonne = 1000 kg) and carried a 1-tonne
warhead as part of its 4-tonne throw-weight. It was fueled with liquid oxygen and
ethanol with Vex = 2 km/s. The burn-out velocity from the rocket equation is
vf = Vex ln(mo /mf ) = (2 km/s) ln(12.8 tonne/4 tonne) = 2.3 km/s, (2.6)
which must be reduced to account for gravity, air drag, and inefficiencies.
2.1.2 SS-18
The SS-18 is a huge, two-stage rocket, which delivers 10 warheads 11,000 km with an
8-ton throw-weight. The velocity gained by the two-stage system (gravity-free) is
v = vf = Vex ln[(m1 + m2 + mTwt )/(m1s + m2 + mTwt )]
+ Vex ln[(m2 + mTwt )/(m2s + mTwt ]. (2.7)
P1: GFZ
2.2. ICBM Trajectories 33
The SS-18 parameters used by the APS study are as follows: First stage mass
m1 = 146.2 tonne, second stage mass m2 = 30.4 tonne, throw-weight mTwt = 8 tonne,
exhaust gas velocity Vex = 3.0 km/s, and the mass of an empty stage is 13% of its
initial mass, or m1s = 0.13 m1 . Using these values gives a final velocity,
vf = 3.5 km/s + 3.5 km/s = 7 km/s, (2.8)
which is close to the actual value. The two stages contribute equally to the final
velocity. Note that the final velocity would be considerably less if the mass of the
two stages were contained in a single stage,
vf = Vex ln[184.6 tonne/(19.0 + 4.0 + 8)(tonne)] = 5.4 km/s. (2.9)
The launch-weight to throw-weight ratio is
mLwt /mTwt = (146.2 + 30.4 + 8)(tonne)/8 tonne = 23. (2.10)
2.2 ICBM Trajectories

2.2.1 Flat Earth
To avoid complications to calculations of elliptical orbits, we assume parabolic
trajectories above a flat Earth with gravity. This understates missile range because
gravity is reduced to 64% at a 1000-mile altitude (4000 mi/5000 mi)2 . In addition
a round-Earth horizon continually drops, making round-Earth trajectories longer.
This discrepancy is apparent using a launch velocity of 8 km/s. The flat-Earth
solution gives a too-short range of 6400 km, while the spherical-Earth solution is
an elliptical orbit.
To obtain realistic answers for the flat Earth, the launch velocity is overstated
at 10 km/s, exceeding 7 km/s for ICBMs and 7.5 km/s for low-Earth orbits. The
minimum energy trajectory for the flat Earth is realized when missiles are aimed
45◦ above the horizontal, while the minimum energy trajectory of the round-Earth
is 22◦ above the horizontal for a 10,000-km range. We will launch our missile at 30◦
for our calculation, primarily to avoid problems with the tangent function when
considering accuracy. The parabolic range for the flat Earth is
X = v2 sin(2θ )/g = (104 m/s)2 sin 60◦ /(9.8 m/s2 ) = 8800 km, (2.11)
and its flight time is
t = 2v sin 30◦ = 103 s = 20 min, (2.12)
close to the 30 min it takes an ICBM to travel 10,000 km.
2.2.2 Spherical Earth

Elliptical orbits are relatively easy to use, but calculations for optimum launch
angle, time-of-flight and error coefficients are complex. Trajectories are obtained
P1: GFZ
from
d 2r/dt 2 − r (dθ/dt)2 = −G M/r 2 (2.13)
d/dt(r 2 dθ/dt) = 0. (2.14)
These equations in r and θ are relatively easy to solve with Runge–Kutta routines,
which allow corrections for variable thrust, drag force, the nonspherical Earth, and
so forth. For the spherical Earth, equations in x and y are easier to solve (RE is the
Earth’s radius):
d 2 x/dt 2 = −x(g R2E )/(x 2 + y2 )3/2 (2.15)
d 2 y/dt 2 = −y(g R2E )/(x 2 + y2 )3/2 . (2.16)
2.3 ICBM Accuracy

The reported accuracy of US Peacekeeper and Trident II systems is 90–100 m from a
target some 10,000 km away, due to errors in initial parameters of 10 parts per mil-
lion (10 ppm). The accuracy of a ballistic missile is determined from the following
errors:
(1) terminal velocity v = 0.5 × 10−5 v = 0.5 × 10−5 (104 m/s) = 0.05 m/s
(2) range, vertical angular error 10−5 θ = (10−5 )(0.5 rad) = 0.5 × 10−5 rad
(3) tracking azimuthal error φ = 10−5 rad.
The range error X is the product of the fractional range error (X/X) times the
range X:
X = [2(v/v) + 2[θ/tan(2θ )] + (g/g)]X. (2.17)
Using the above values, the range error over the 8800-km range from the velocity
error v/v is
X = 2(v/v)X = [2(0.05/104 )](8.8 × 106 m) = 88 m. (2.18)
The range error from error in the vertical angle θ is a second-order correction
when launching at the minimum energy angle of θ min = 22◦ above the horizon
for 10,000-km range flights above round-Earth. Since we are not at that angle, the
range error is
X = 2[θ/tan(2θ )]X = [2(0.5 × 10−5 rad)/tan60◦ ](8.8 × 106 m) = 51 m. (2.19)
If these are random errors, the combined range error is
σx = (882 + 512 )1/2 = 102 m. (2.20)
If the errors were systematic miscalibrations, the total error could be as large as
139 m. The error in azimuthal angle φ gives rise to an error in the tracking direc-
tion,
Y = (φ)X = (10−5 )(8.8 × 106 m) = 88 m. (2.21)
P1: GFZ
2.3. ICBM Accuracy 35
The radial error from the aim point is obtained by combining the range and tracking
errors,
σtotal = (1012 + 882 )1/2 = 130 m. (2.22)
2.3.1 Rotating Earth

Coriolis and centripetal accelerations must be taken into account when the missile
is aimed at Earth-based coordinates. Indeed, ICBM inertial guidance systems take
into account acceleration, the force of gravity, and the moving positions of the
launch site and the target site. For instance, during an ICBM flight of time T =
30 minutes, a site at β = 45◦ north latitude moves to the east
x = ω RE (cos β)T = (7.29 × 10−5 rad/s)(6.4 × 106 m)(0.707)(1800 s) = 600 km.
(2.23)
An error of 10 ppm in flight time would increase the aiming error by 5%.
ICBM errors are caused by the following factors:
r initial ballistic velocity and direction
r accelerometer (bias, calibration, misalignment, vibrations)
r gyroscope (initial and acceleration induced drift, vibrations)
r thrust termination
r energy loss maneuvers
r gravitational anomalies
r guidance computation
r reentry buffeting and fusing.
US accuracy improved from 1400 m in 1962 to 90 m in 1988. The Soviets were on
average about 7 years behind the US trend, improving from 2000 m in 1961 to 230 m
in 1986. These errors due to Earth’s rotation can be addressed by updating three-
dimensional position and velocity vectors during flight via star locations. Since the
kill probability (with reliability = 1) from attacks by hard-target warheads on silos
is close to 1.0, there is no great need to increase accuracy for those cases. It would be
possible to develop accuracy of 50 m through maneuvering reentry vehicle (MaRV)
technology, which the United States used in the US Pershing II, but this approach is
very expensive. Better accuracy could be used to lower weapon yields for attacking
fixed points, but such accuracy is already available with cruise missiles.
2.3.2 Nonspherical Earth Gravitational Bias

US and Soviet-Russian ICBMs are intended to travel near the North Pole. Because
Earth’s polar radius is 21 km (0.3%) smaller than its equatorial radius, guidance
computers must take into account the nonspherical Earth. Highly accurate three-
dimensional, gravitational multipole-potentials were developed for the Earth by
observing variations in satellite orbits. When a satellite approaches a concentrated
extra mass, the satellite speeds up slightly and it slows after it passes the mass
P1: GFZ
concentration. Corrections for local gravity at launch sites are important, since
slowly rising missiles spend more time near the modified gravitational force.
We consider only the quadrapole term with a simplified approach that uses g/g
to determine the gravitational bias error. To first order, the fractional change in g is
proportional to the fractional change in Earth radius, or
g/g ≈ RE /RE = 0.003. (2.24)
◦
Because missiles take off and land at about 40 north latitude, far from the equator,
the estimate of the bias error is reduced by a factor of about 3; that is
X = (g/3g)X = (10−3 )(8.8 × 106 m) ≈ 15 km, (2.25)
which agrees with accurate estimates. Guidance computers must calculate gravi-
tational bias corrections to better than 1% accuracy because a 15-km error is 100
times larger than 100-m accuracy. The conventional wisdom is that good guidance
computers can do this calculation.
2.4 GPS Accuracy

US global positioning satellites (GPS) and Russian global navigation satellites
(GLONAS) can be used to accurately determine locations at receivers on and above
the Earth. Cruise missiles can passively receive GPS signals to determine its loca-
tion to a few meters using the unclassified channel. GPS navigation is not detectable
since GPS signals are passively received; this is not the case for cruise missiles that
send radar signals, which can be detected by the other side. The absolute location
of a cruise missile can be determined to less than a meter by using the unclassified
channel and referencing its location to known locations, using differential GPS nav-
igation. Scientists have devised ways to circumvent degraded unclassified GPS
signals to achieve results better than expected. GPS is helpful to ground troops,
ships, bus drivers, airlines, surveyors, hunters, and cruise missiles.
2.4.1 GPS Triangulation

Position is obtained by triangulation from the timed signals from 3 or more of the
24 GPS satellites, spaced at 15◦ intervals at a radius R = 24,000 km from Earth’s
center. To simplify our estimates we will assume that all the missiles and the GPS
are in the equatorial plane. Assume GPS receivers measure time to better than 1 ns
using hydrogen maser atomic clocks accurate to 10−13 s. For simplicity, consider a
cruise missile directly under a GPS satellite in the zenith position on the equator.
Signals from the first-neighbor satellites, 15◦ on either side of the zenith satellite,
arrive at the same time.
The movement of a cruise missile is obtained by taking the differential of the
law of cosines. Chord A is between the zenith satellite and its first neighbor, side
B is between the zenith satellite and the cruise missile (17,700 km), and side C is
between the first-neighbor and the cruise missile. The chord length is (Fig. 2.1)
A = 2R sin 7.5◦ = 2(24,150 km)(0.13053) = 6304 km. (2.26)
P1: GFZ
2.4. GPS Accuracy 37
Figure 2.1. Global positioning satel-

lites. As a missile moves ahead 1 m,
side C increases by 0.35 m with an angle
ε = 5.6 × 10−8 rad. This increases the
transit time by a measurable 2.4 ns.
Since the lengths A and B are known and the outside angle between these sides is
105◦ , the distance between the first neighbor satellites and the cruise missile is
C = [A2 + B 2 + 2AB cos 105◦ ]1/2 = 17,184 km. (2.27)
When the cruise missile moves forward 1 m from its position directly under the
zenith GPS, there is a delay time separating pulses from the two first-neighbor
global positioning satellites. Since a 1-m shift increases distance B by only one part
in 1014 , we consider B and A to be constant. The change in C is mainly caused by
a change in angle ε between A and B, which is
ε ≈ tan ε = 0.001 km/17,700 km = 5.6 × 10−8 . (2.28)
The distances from the two nearest-neighbor satellites to the cruise missile could
be determined from the law of cosines by adding (and subtracting) angle ε from
the 105◦ angle. But it is easier to take the differential of the law of cosines, keeping
A and B constant, to get the additional path length,
C = Cf − Ci = ABε sin 105◦ /C = 0.35 m. (2.29)
The delay time between pulses from the two nearest-neighbor GPS satellites is
t = (2 × 0.35 m)/(3 × 108 m/s) = 2.4 ns. (2.30)
One-meter accuracy is obtained since the 2.4 ns delay for 1-m displacement is
readily measured. Accuracy of a few millimeters can be obtained from 1% accurate
measurements of the phase shift between two 19-cm waves (1.5 GHz).
P1: GFZ
2.5 Kill Probability = f (CEP, H, Y, R, n, Fratricide)

After the two superpowers spent $10 trillion to build 100,000 nuclear weapons
and weapon delivery systems, one might ask the retrospective question, “How
much was enough?” The path to an answer should begin with determining param-
eters and performing an analysis, and then continue with a discussion of political,
theological, sociological, psychological, and historical implications. The survival
of missile silos depends on three basic parameters (CEP, H, and Y): (1) Accuracy
of missiles given in circular error probable (CEP) radius of a circle in which 50% of
warheads fall; (2) maximum overpressure that a silo survives, called hardness (H).
What actually destroys a silo is the delivered impulse (force ×time), but this is
simplified here to hardness. US silos are hardened to about 2000 psi, while the Sovi-
ets built some silos to take larger overpressures. Hardening is increased with better
construction, reduced coupling with hanging straps to hold ICBMs, and massive
springs and shock absorbers; (3) yield (Y) of weapons, given in kilotons (kton) or
megatons (Mton). Two more parameters must be considered: (4) The reliability (R)
of a weapon varies between 0 and 1 and (5) the number (n) of warheads targeted at a
silo. Each successive warhead gives a smaller additional probability of destroying
the target because each previous warhead may have already destroyed the tar-
get. The functional kill probability for n attacking warheads uses five parameters,
Pkill-n = f (CEP, H, Y, R, n).
The “cookie cutter” approximation assumes that a target is destroyed if overpres-
sure exceeds the hardness of the target but it survives if the overpressure is less
than the hardness. Reality expects that the step function, cookie cutter probability,
which is either 0 or 1, should be smoothed with a function similar to the Fermi-Dirac
distribution function. However, the additional sophistication of a smoothed cookie
cutter does not change the results because uncertainties in parameters are larger
than analytical gains of adding another parameter. A two-dimensional Gaussian
kill probability density fucntion describes missiles impacting a distance r from a
target,
p(r ) = (1/2π σ 2 ) exp(−r 2 /2σ 2 ). (2.31)
In practice, the footprint of landing missiles is an ellipse, but we will treat it as a

circle. If gravitational bias exists, r should be replaced by the vector (r − B). Missile
accuracy is not quoted with a standard deviation σ s but with CEP radius. By inte-
grating p(r ) from r = 0 to CEP and setting the integrated single-shot kill probability
(SSKP) to 0.5, it is shown in problem 2.13 that CEP = 1.1σ . The general form of
the equations below can be discovered with little physics by using dimensional
analysis in problem 2.11. By using the radial dependence of overpressure and the
definition of CEP, the SSKP (reliability = 1) is
SSKP = 1 − exp(−Y2/3 /0.22H 2/3 CEP2 ), (2.32)
where Y is in megaton, H is in psi, and CEP is in nautical miles (1860 m). The kill
probability for one warhead takes into account reliability of the missile-warhead
P1: GFZ
2.5. Kill Probability = f (CEP, H, Y, R, n, Fratricide) 39
system,
Pkill-1 = R × SSKP. (2.33)
2.5.1 Accuracy Versus Yield

A 1-Mton warhead has an SSKP of 90% against a target. By how much can yield be
reduced if accuracy is improved by a factor of 2, while retaining the same SSKP?
Using a fixed SSKP argument with H1 = H2 gives
(Y1 /Y2 )2/3 = (CEP1 /CEP2 )2 and (Y1 /Y2 ) = (CEP1 /CEP2 )3 . (2.34)
Thus, a CEP reduced by a factor of 2 allows the yield to be reduced by a factor of

8. For our example, this gives a reduced yield of (1/2)3 (1 Mton) = 1/8 Mton. US
weapon yield was reduced as accuracy was improved by a factor of 4 as Minuteman
II (0.2 nmi = 370 m) was replaced with Peacekeeper (0.05 nmi = 90 m). The reduction
by a factor of 4 in CEP implies that yield could be reduced by 43 = 64, but yield was
in fact reduced only by a factor of 4 from Minuteman-II to Peacekeeper. The cause
of the difference between ratios of 4 and 64 is that the Peacekeeper was designed
for harder silos and in an era when higher kill probabilities were sought.
The Soviets always had larger weapons because Soviet accuracy was always
surpassed by the United States. Even today, the reported accuracy of the SS-18
(0.13 nmi) is about 1/3 that of Peacekeeper’s 0.05 nmi. In Senate hearings on the
Strategic Arms Limitation Treaty (SALT), much was made of the large size of Soviet
SS-9s as compared to US Minuteman. Senators misled the public by showing large
models of Soviet missiles emphasizing launch-weight and yield, but they neglected
the two most important parameters, accuracy and reliability.
2.5.2 Accuracy Versus Hardness

As US accuracy increased, the Soviets moved their ICBMs from launch pads to
silos with 300-psi hardness, then to silos with 2000-psi hardness and finally to a
few silos with greater hardness. During this period, US accuracy improved from
1300 m in 1962 to 300 m in 1970 to 90 m in 1986. It is generally accepted that
accuracy won the race against hardness. Perhaps superhardened silos might be
able to withstand 10,000 psi, but the cost would become very large. In addition,
when crater size becomes similar to CEP, the kill mechanism becomes cratering,
and not overpressure. US hard-target warheads can produce craters with radii
approaching their accuracy.
2.5.3 Relative Constancy of Hard-Target Yield

The United States and Russia maintain warheads of about one-half Mton for their
hard-target weapons. The record was set with the test of the Soviet’s 58-Mton
weapon in 1962, which was later reported to be but a part of a 100–150 Mton
P1: GFZ
weapon. These parameters are consensus numbers from the International Institute
for Strategic Studies.
2.5.4 Reliability Versus SSKP

Warhead accuracy requires many tasks be carried out reliably. The total reliabil-
ity of a ballistic missile is the product of the reliabilities for command-control-
communication-intelligence (C3 I) reliability, missile reliability, and warhead relia-
bility:
Rtotal = RC3I × Rmissile × Rwarhead . (2.35)
The US Congressional Budget Office quoted a reliability of 85% for US ICBMs.
It is generally believed that warheads have a high reliability of greater than 95%,
higher than the reliability of the missiles that carry them. The ratio of missile-to-
warhead failure rates (F ) is, perhaps, a factor of 3, from these reliabilities:
Fmissile /Fwarhead = (1 − Rmissile )/(1 − Rwarhead ) = (1 − 0.85)/(1 − 0.95) = 3. (2.36)
Consider the case where high-yield, accurate missiles have SSKP ≈ 1. The sur-
vival probability is 1 − Pkill-1 = 1 − R. For hard-target weapons (Peacekeeper, Tri-
dent/W88, SS-18, SS-27), the number of surviving targets is essentially determined
by the reliability of the attacking system.
2.5.5 Lethality
A warhead’s prowess is discussed in terms of its main parameters of yield and
accuracy by combining it into the lethality (L) parameter,
L = Y2/3 /CEP2 . (2.37)
Note that L is proportional to the ratio of the destroyed area (proportional to Y2/3 )
to the missile arrival area (CEP2 ). This ratio appears in the exponential argument for
SSKP. In debates on the arms race, prestigious individuals have compared the total
lethality of the two superpowers to determine which side was ahead in the arms
race. This method employs poor logic since L does not take into account the hard-
ness of targets. Also a missile with a tremendous L value (very accurate with very
large yield) could have essentially the same kill probability as a weapon which has
only 50% as much lethality, since the exponential term in kill probability would be
saturated. Lethality is useful as a starting point, but it is only a beginning.
2.5.6 Rate of Change in Pkill-1

Parameter changes affect kill probabilities. It is useful to take the differential of the
single warhead kill probability Pkill-1 :
Pkill-1 = R(1 − e −α ), (2.38)
P1: GFZ
2.6. Nuclear Conflicts 41
Table 2.1. Improvements from enhanced R, Y, H, and CEP. Improvements in one-warhead kill
probability, Pkill-1 /Pkill-1 , from 10% improved reliability, yield, hardness, and accuracy for
two situations.
Attacker Y(Mt) H(psi) CEP(nmi) R(0–1) L Pk1 Pk2 Pk1 /Pk1 (%): R Y or H CEP
A 0.75 2000 0.135 0.85 45 62% 85% 10% 3.3% 9.8%
B 0.5 2000 0.05 0.9 252 89.9% 99% 10% 0.04% 0.1%
where α = Y2/3 /(0.22 CEP2 H 2/3 ) to obtain
Pkill-1 /Pkill-1 = R/R + (2α/3)[Y/Y − H/H − 3(CEP/CEP)]/(e α − 1). (2.39)
It is clear that situation B is much better than situation A. It takes two A-warheads
to accomplish what B can do with one. It follows that A improves its one-warhead
kill probability more with 10% improvements than 10% improvements for B. For
A, 10% improvements in reliability (R/R = 0.1) and accuracy (CEP/CEP = 0.1)
gives 10% improvements in Pkill-1 , while a 10% yield increase (Y/Y = 0.1) raises
Pkill-1 by 3.3%. For B, which has much better accuracy, 10% improvement increase
Pkill-1 by 10% for reliability, 0.04% for yield and 0.1% for CEP (Table 2.1).
2.6 Nuclear Conflicts

The most likely nuclear conflict would probably start as an accident, driven by
false information, rather than a sudden “bolt out of the blue.” In 1979 missile crews
received warning that a massive nuclear attack from the Soviet Union was under
way. Luckily, President Carter decided not to respond with a launch-on-warning
attack. Later it was discovered that the attack was bogus, the result of a training
tape accidentally left in the computer system. A similar crisis took place in 1995
when a Black Brant rocket, launched from Norway, was interpreted by the Russians
as a Trident attack. Fortunately, the nuclear response was averted when Russian
President Boris Yeltsin ignored the false alarm. It is for this reason that many have
called for placing additional strategic weapons on an off-alert or delayed-alert
status. Nuclear policy is based primarily on worst-case analysis, that is when one
side covertly surprises the other side with an all-out attack. The launch on warning
scenario is driven by the fear that it is better to use all the weapons on a timely basis,
rather than have fewer weapons aimed at empty silos. These bolt-out-of-the-blue
scenarios guide the calculations of this section and Section 4.11.
2.6.1 Two Warheads Per Target

The survival probability for one warhead attacking a silo is
Psurvive-1 = 1 − Pkill-1 = 1 − R × SSKP. (2.40)

P1: GFZ
If two warheads with the same parameters, but coming from different missiles,
attack a silo, the survival probability is multiplicative, since the launches are inde-
pendent actions,
Psurvive-2 = (1 − R × SSKP)(1 − R × SSKP) = (1 − R × SSKP)2 , (2.41)
with a total kill probability of
Pkill-2 = 1 − (1 − R × SSKP)2 . (2.42)
Since missile reliability is the most likely failure mode, warheads from different
missiles are used to target a silo. A failure of a missile carrying two warheads for
one target would cause both the first and second warheads to fail. For the case of
SSKP = 1 and Rmissile = 0.8, 80% of the silos would be destroyed and 20% would
survive, since second warheads fail with the first failure. If different missiles were
used for the two warheads, the kill probability would be raised to 96%:
Pkill-2 = 1 − (1 − 0.8)2 = 1 − (0.2)2 = 96% and Psurvive-2 = 4%. (2.43)
2.6.2 Fratricide
The equation for Pkill-2 assumes the two warheads are truly independent. But, if
both explosions are surface blasts (or both high-altitude air bursts), one warhead
might miss the target, and still destroy the second warhead. Kill probability for two
warheads might be increased by using one surface warhead and one high-altitude
warhead, but the high altitude blast would not destroy the silo. There are many
mechanisms that cause fratricide, the killing of one warhead by another: Blast waves
and dust can destroy the second warhead; an electromagnetic pulse from the first
warhead can destroy the second warhead’s electronics; and neutrons from the first
warhead can preheat or pre-initiate the second warhead. Most of these effects take
place in a narrow time window, reducing the problem, but dust from a first surface
blast can damage the second surface blast warhead. The timing separation needed
between explosions is difficult to obtain, considering that the two warheads are
launched from separate missiles 10,000 km away.
For the case of no fratricide,
Pkill-2 = 1 − (1 − R × SSKP)2 = (2R × SSKP) − (R2 × SSKP2 ). (2.44)
We consider three fratricide situations: (1) The first warhead destroys the target
with a probability of R× SSKP. (2) The first warhead misses the target, but destroys
the second warhead with reliability R. (3) The first warhead misses the target, but
does not destroy the second warhead. For simplicity, we consider completely effective
fratricide and ignore the third possibility to obtain (problem 2.17),
Pkill-2-fratricide = (2R × SSKP) − (R2 × SSKP). (2.45)
This result slightly differs from Pkill-2 . For very reliable and lethal weapons (R = 1,
SSKP = 1), fratricide is irrelevant since it takes only one reliable warhead to destroy
a silo. However, if reliability is not 1, but the weapons are very lethal with SSKP = 1,
P1: GFZ
Figure 2.2. Fratricide. The

number of silos that survive
calculated as a function of
accuracy for two situations:
(a) no fratricide and (b) totally
effective fratricide when a first
warhead misses a target and
destroys a second warhead.
Totally effective fratricide
increases the number of
surviving silos by about 10%.
then two-shot kill probability with fratricide and without fratricide reduces to the
same answer, Pkill-2-fratricide = 2R – R2 = Pkill-2 . However, when SSKP is not 1 there
is a marked difference. In Fig. 2.2, we plot, as an example, the number of surviving
silos as a function of accuracy using the above equations for Pkill-2 and Pkill-2-fratricide .
Accuracy is varied while yield and reliability remain fixed. The curves with and
without fratricide coincide for accuracy better than 0.06 nautical mile since SSKP
approaches 1 at that point, but they separate for larger CEPs. The shaded area
indicates that, at most, 100 additional silos (10% of 1000) could survive because of
fratricide.
2.6.3 More Than Two Warheads Per Target

By simple extension, the n-shot kill probabilities are
Pkill-n = 1 − (1 − R × SSKP)n = 1 − (1 − Pkill-1 )n . (2.46)
It does not make great technical sense to use a third warhead on a target when
the marginal return is small or when fratricide is increased. For the case of good
hard-target weapons, Pkill-1 is about 0.9. A second warhead makes an improvement
to Pkill-2 = 0.99, but a third warhead gives only marginal improvement at Pkill-3 =
0.999. This argument is weakened if Pkill-1 is low, say Pkill-1 = 0.5, giving Pkill-2 = 0.75
and Pkill-3 = 0.875. In this case, improvements with each additional warhead is
larger. Government calculations using conjectured three-warhead targeting by the
P1: GFZ
Soviets were used in Senate testimony by those trying to show vulnerability of US

systems.
2.6.4 Earth-Penetrating Warheads

In a crisis, leadership and command-and-control authority would retreat into hard-
ened, underground bunkers. Such bunkers could also store nuclear or biological
weapons. Weapons explosions coupled through air overpressure to the ground
might not be sufficient to destroy deep underground bunkers. However, coupling
would be greatly enhanced if a weapon penetrated the Earth and exploded, di-
rectly coupling to the ground without the intermediary air. It is for this reason that
the United States developed the B61/11 earth-penetrating warhead, which magni-
fies pressures by a factor of 30. For this to take place, earth-penetrating warheads
must maintain their physical integrity while burrowing through several meters of
earth. The 2002 Nuclear Posture Review recommended studies on 5-kton penetrat-
ing weapons. It has been conjectured that a 5-kton weapon could be a clean bomb
because it could bury its radioactivity debris. However, this is not true because
the weapon’s crater size is greater than its penetration depth. For this reason, the
congress did not support the robust earth penetrating warhead.
2.6.5 Minuteman and SS-18 Vulnerability

Concern has been voiced over the years that ground-based ICBMs are vulnerable.
In spite of this concern, Peacekeeper (PK) was based in silos that were vulnerable,
which is one of the reasons it has been withdrawn. Silo vulnerability has to be dis-
cussed within the context of the other two legs of the triad, the heavy bomber force
(B-52, B-1, and B-2) and the Trident submarines. In Section 4.11, we estimate nu-
clear survivability under the now discarded START II. The issue of silo vulnerability
would be less important if MIRVed systems were discarded. With fewer warheads,
two-on-one targeting becomes too costly to consider. Mathematics might imply the
winner would be the side with the most remaining warheads, but such a deduction
would ignore the other triad legs and the vulnerability of cities.
2.6.6 MIRV and Stability

PK was finally based in former Minuteman silos due to a lack of viable options.
It was understood that this basing was unstable. It is possible for 10 SS-18s with
100 warheads to attack 50 silos, destroying 90% of 500 warheads on 50-PK missiles
in a worst possible case scenario. The ratio of warheads destroyed/used for the case
of the Soviets attacking first would be (US destroyed)/(USSR attack) = 450/100 =
4.5. This situation would be reversed if PK were to attack first, reversing the ratio to
0.22. The ratio of ratios determined by which side were to attack first is 4.5/0.22 =
20. In spite of this large instability, PK was placed in vulnerable silos, knowing that
other US warheads would deter attacks. But this approach magnified the potential
P1: GFZ
problem of protecting the PK with launch-on-warning preemption that could be

based on incorrect information.
2.6.7 Other Basing Modes

There are no completely survivable silos. This shortcoming is not too serious for
single-warhead ICBMs, since it takes two warheads to reliably destroy a single
warhead in a silo, making these silos “sinks” for warheads. On the other hand,
SS-18s and PKs have 10 reentry vehicles per missile, which makes them strategi-
cally vulnerable. In 1981, the Office of Technology Assessment examined 10 basing
modes for PK to overcome this vulnerability, but they failed to identify a viable
mode among the following options.
r 200 PK moved into 4600 horizontal/vertical shelters with decoys
r shelters/silos with ABM defense
r launch-under-attack, launch-on-warning
r small submarines or surface ships, each with 2 PK
r 75 wide-bodied aircraft with dash-on-warning
r densely packed silos defended by fratricide
r deep underground basing, PK burrow out after attack.
2.6.8 Comprehensive Test Ban Treaty

The Comprehensive Test Ban Treaty (CTBT) would ban nuclear tests of any yield
in all places for all time. The CTBT is an arms control measure that constrains
the five nuclear weapons states from developing new weapons. The United States
has tested 1030 times, much more than China (45), the United Kingdom (45), and
France (210). Russia has tested 715 times, but its situation has made it more diffi-
cult to develop new strategic weapons. The CTBT is also a nonproliferation mea-
sure since a test ban raises a finite barrier to development of first-time nuclear
weapons. The 1998 tests by India and Pakistan highlighted the need for a CTBT
(Fig. 2.3).
The CTBT has 176 signatories (July 2006), which amounts to practically all the
nuclear capable nations, except for India, Iraq, North Korea, and Pakistan. Non-
nuclear weapons states (NNWS) view the CTBT as the “quid pro quo” by nuclear
weapons states (NWS) to the nonproliferation regime (Chapter 5). In October 1999,
the US Senate rejected the CTBT with a vote of 51 to 48. After the defeat, the Na-
tional Academy of Sciences was asked by the Clinton administration to convene a
panel of experts to examine technical issues that may affect the viability of a test
ban (National Academy of Sciences, 2002). The panel looked at
r US capacity to maintain safety, reliability, and design of its nuclear stockpile
without testing;
r ability to monitor a test ban, including evasion scenarios (Sections 4.8–4.10);
r ability of nations to increase nuclear capability by cheating and the effect on US
security.
P1: GFZ
Figure 2.3. Worldwide nuclear tests (1945–2002). The United States tested 1030 times and
the former Soviet Union tested 715 times. The third curve combines the 306 tests of the
United Kingdom (45), France (210), China, India (4), and Pakistan (2). [Data from R. Norris
and W. Arkin, Bulletin of Atomic Scientists 54, 65–67, November/December 1998]
The Academy panel stated that several factors were key in its analysis on the abil-
ity of the US stockpile stewardship program to maintain the safety and reliability
of its nuclear weapons without testing. These were as follows:
r Confidence will require a high-quality workforce and adequate budgets.

r Stockpile stewardship and enhanced surveillance must examine nuclear compo-
nents.
r Remanufacture to original specifications is the preferred remedy for age-related
defects.
r Primary yield that falls below the minimum level needed to drive a secondary is
the most likely potential source of nuclear-related degradation.
r Based on past experience, the majority of aging problems will be found in the non-
nuclear components, which can be fully tested under a CTBT. (NNSA has stated
that the nuclear Pu pits have a minimum lifetime of about 100 years (2006) with
“no life-limiting factors.”)
r A highly disciplined process is needed to install changes in nuclear designs.
r In the past, confidence tests were limited to one per year, as most tests were
carried out to critique new designs.
P1: GFZ
From these results, the Academy panel concluded the following:

“It seems to us that the argument to the contrary—that is, the argument that improvements
in the capabilities that underpin confidence in the absence of nuclear testing will inevitably
lose the race with the growing needs from an aging stockpile—underestimates the current
capability for stockpile stewardships, underestimates the effects of current and likely future
rates of progress in improving these capability, and overestimates the role that nuclear testing
ever played (or would ever by likely to play) in ensuring stockpile reliability.”
The US secretaries of Defense and Energy are required to make an annual certifi-
cation on the status of the nuclear stockpile and whether the stockpile stewardship
program is maintaining the warheads without testing. One approach to ensuring
maintenance would be to require a fixed reliability level. Another approach would
be to determine how much reliability is enough for various warhead missions,
such as the following: (1) Nuclear warheads could be used to respond to an attack
by Russia or China. (2) They could be used to attack Russia or China first. (3).
They could be used to respond to an attack by a smaller nation. (4) They could be
used to threaten or attack a smaller nation first. Reliability for option (1) is of less
importance since many silos are empty and cities are soft. Option (2) needs high
reliability weapons to attack strategic targets first. Reliability for options (3) and
(4) is of less importance because there are many more US weapons available to
overtarget. Ultimately, the main purpose for nuclear weapons is deterrence, but
one also has to consider the targets.
2.6.9 Nuclear Triad

Was the nuclear triad of ICBMS, SLBMs, and heavy bombers vulnerable in attack,
requiring the full triad to truly deter?2 Or, was the Cold War driven by consideration
of a “worst-case” analysis? The 1992 US General Accounting Office (GAO) report
concluded that the Soviet threat to the US triad was overstated, that the performance
of existing US systems was understated, and that the performance of new US
systems was overstated (Fig. 2.4).
2.6.10 ICBMs
The GAO concluded that
“In the case of the land leg, . . . the claimed ‘window of vulnerability’ caused by improved
Soviet missile capability against [US] silo-based ICBMs was overstated on three counts. First,
it did not recognize the existence of sea and air leg deterrence—that is, the likelihood that
the Soviets would hesitate to launch an all-out attack on the ICBM silos, given their inability
to target submerged US submarines or on-alert bombers and their thousands of warheads
that could be expected to retaliate. Second, the logic behind the claim assumed the highest
estimates for Soviet missile performance dimensions as accuracy, yield and reliability, while
2
The 2002 Nuclear Posture Review defined the new triad as consisting of nuclear weapons,
conventional weapons, and an information-based transformed military.
P1: GFZ
Figure 2.4. 1984 Draw-down curve. The number of US warheads that survive a Soviet first-
strike is estimated as a function of the number of attacking Soviet warheads. The first section
of the curve denotes the reduction in bombers and submarines in port. The second section
denotes the reduction in silo-based warheads from a single-warhead attack on each silo. The
third section denotes the reduction in silo-based warheads when two Soviet warheads are
targeted on each silo. Note that the marginal return from the use of additional warheads
decreases rapidly.
at the same time discounting very substantial uncertainties about performance that could not
have been resolved short of nuclear conflict. Third, it ignored the ability of US early warning
systems to detect a Soviet ICBM attack, and thereby allow a reasonably rapid response.”
[The US Nuclear Triad, US General Accounting Office, 1993]
2.6.11 Submarine-Launched Ballistic Missiles

In a similar fashion, GAO concluded that threats to SLBMs had been overstated in
“unsubstantiated allegations about likely future breakthroughs in Soviet subma-
rine detection technologies, along with the underestimation of the performance and
capabilities of [US] nuclear powered ballistic missile submarines.” The threats to
the SLBMs have been categorized as “non-acoustic anti-submarine warfare,” which
use radar, laser, or infrared detectors on satellites to search out the signatures of
the nuclear submarines. Two submarine signatures that have been identified are
the slightly raised ocean surface above a moving submarine (the Bernoulli hump)
and the V-shaped wave above a moving submarine (the Kelvin wave). In principle,
P1: GFZ
these signatures might be observed from submarines near the surface if one were
to know where to look, using orbit-based synthetic-aperture radar accompanied
by significant computer capabilities, but this would be very difficult.
2.6.12 ICBMs Versus SLBMs

GAO concluded that the offensive power of the sea leg (SLBM) was essentially
equivalent to that of the land leg (ICBM): “The sea-leg’s performance has been
understated (or poorly understood) on a number of critical dimensions. Test and
operational patrol data show that the speed and reliability of day-to-day communi-
cations to submerged, deployed SSBNs [ballistic missile submarines] are far better
than widely believed, and about the equal in speed and reliability of communica-
tions to ICBM silos. Yet conventional wisdom gives much higher marks to ICBM
command and control responsiveness than to that of submarines. In point of fact,
SSBNs are in essentially constant communication with national command author-
ities and, depending on the scenario, SLBMs from submarine platforms would be
almost as prompt as ICBMs in hitting enemy targets. Other test data show that
the accuracy and reliability of the Navy’s D-5 SLBM are about equal to DOD’s
best estimates for the Peacekeeper. Further, its warhead has a higher yield than the
Peacekeeper’s. In short we estimate that the D-5 has a hard target kill capability
about equal to the Peacekeeper’s, while its platforms remain virtually undetectable,
unlike easily located silos.”
2.6.13 Breakout from SORT

In 1990, the two superpowers declared that they each had about 12,000 deployed
strategic nuclear warheads. The 1991 START I lowered this to 7000–8000 warheads.
Both the US Congress (January 1996) and the Russian Duma (April 2000) ratified
START II (3500 warheads, ban on MIRVed ICBMs), but the treaty did not enter into
force because President George W. Bush withdrew the United States from the ABM
Treaty (June 2002). A limit of 2000–2500 warheads was agreed to by President Bill
Clinton and President Boris Yeltsin in Helsinki in March 1997 for a prospective
START III, but the Congress and the Duma blocked progress. On March 6, 2003,
the US Senate ratified the Strategic Offensive Reduction Treaty (SORT) with a limit
of 1700–2200 “operational” warheads for 2012. This limit is the same as the START
III limit since it ignores 240 warheads on two Trident submarines in overhaul. The
2002 Nuclear Posture Review used 2200 operational warheads on Minuteman (450),
operational SLBMs (1440), and bombers (300). When added to 1000 warheads on
additional bomber positions, plus 3000 more for the hedge, reserves, and tactical
missions, the total could rise to about 6000 warheads. Russia protested that they
wanted more transparency on warhead reductions because of these large numbers,
but in the end they accepted this approach without further verification measures
because they could retain 138 SS-18s and a considerable number of SS-19s.
The main concern regarding a nation cheating on treaties at lower levels of war-
heads is the possibility that downloaded MIRVed missiles (SS-19, Minuteman) or
P1: GFZ
single reentry vehicle missiles (SS-25/27) could be covertly uploaded with ad-
ditional warheads. Russia could upload 1500 warheads (100 SS-19 with 5 more
warheads = 500, plus 500 SS-25/27 with 2 warheads = 1000), while the United
States could upload 3000 warheads (SLBM, 14 subs × 24 SLBM × 3 warheads =
1008, plus 450 Minuteman × 2 warheads = 900, plus more than 1000 on bombers).
See Section 4.11 for a discussion of the verifiability of the START II Treaty.
2.7 Conventional Conflicts

2.7.1 Lanchester Equations
Modeling conventional wars is difficult, as shown by the overly pessimistic Pen-
tagon predictions for the 1991 Gulf War. In 1914, Frederich Lanchester developed
equations to describe destruction of military forces without considering their force
structure and political will to fight. The Lanchester equations assume the x and
y forces are diminished at a rate proportional to the opponent’s strength, which
assumes that all the troops on one side shoot at all the other troops on the other
side:
d Nx /dt = −ε y Ny and d N y /dt = −εx Nx , (2.47)
with fighting efficiency ε x and ε y , and force strengths Nx and Ny . Multiplying these
equations gives
−εx Nx (d Nx /dt) = −ε y Ny (d N y /dt), (2.48)
which integrates to
εx Nx2 − ε y Ny2 = constant. (2.49)
This shows that the effectiveness of a force is the size the force squared times its
fighting efficiency. To simplify matters, assume the sides have equal effectiveness,
ε x = ε y = 1. If the initial forces have a strength ratio of 2/1 with Nx = 2 and Ny =
1, the result of the battle without surrender is
[Nx2 − Ny2 ]initial = 22 − 12 = 3 = [Nx ]2 − 02 , (2.50)
with final forces of Nx = 1.73 and Ny = 0. Everything else being equal, superior
numbers win battles. Side y lost everything, while side x lost only 13% of its forces.
Large initial Nx /Ny ratios give smaller losses for the large side. This computation
can overstate numerical advantages; for example, a smaller side with better tech-
nology willing to fight won at Augincourt, while another small force lost at the
Alamo. The results in Table 2.2 were derived from Eq. 2.50.
Table 2.2. Fractional loss for the larger side x.

x/y force ratio Nx /Ny 1 2 3 5 7
x fractional loss Nx /Nx 100% 13% 6% 2% 1%
P1: GFZ
2.7. Conventional Conflicts 51
2.7.2 The Richardson Model

Lewis Richardson was moved by his World War I experience to research the the-
ory that claims that when a nation acquires weapons in increasing numbers, its
neighbors are encouraged to do the same. This was the case before World War I
when European nations increased their armaments in response to each other, and
then stumbled into war with the “guns of August.” A counterexample occurred
when nations failed to respond to Hitler and the building of his war machine, thus
allowing him to annex his neighbors, resulting in World War II.
There is no simple analysis for all situations, but Richardson showed that military
spending by one side begets military spending on the other side in his study of 315
wars from 1810 to 1953. Richardson developed log–log plots of numbers of conflicts
versus numbers of fatalities. (World Wars I and II had direct military casualties as
high as 3 million per year, but civilian deaths in World War II greatly outstripped
this number.) A conjectured extension of Richardson’s graph gives a probability
of wars having over 10 million fatalities of 1% per year, or one such war in 100
years. World War II was this war in Richardson’s model. An all-out nuclear war
could have killed hundreds of millions, with a Richardson probability of about
0.1% per year (one in 1000 years). At the end of the Cold War, most experts would
downrate the probability of nuclear war. However, the failure to agree on lower
numbers of warheads, dealerting, additional verification measures, the Iraq and
Afghan conflicts and other events are considerations that give one pause.
Richardson’s model gives the action–reaction response of one nation to the threat
of increased military spending by its adversary. The two-nation model uses the
coupled differential equations
d Nx /dt = k N y − a Nx + g and (2.51)

d N y /dt = l Nx − b N y + h (2.52)
to describe the armament levels of two nations, Nx and Ny . First of all, let us consider
only the first terms on the right. The kN y and lNx terms represent the “threat and
response” of a nation to a neighbor’s military spending. Since the k and lcoefficients
are positive, arms races would be exponentially unstable with only the first term.
Adding the equations with only the “threat” terms at k = l gives
d(Nx + Ny )/dt = k(Nx + Ny ), (2.53)
which integrates to give exponential growth,
ln(Nx + Ny ) = kt and (Nx + Ny ) = e kt . (2.54)
Richardson plotted annual increases in armaments, (Nx + Ny )/year, versus

armament level (Nx + Ny ) from 1909 to 1913, obtaining a linear fit as annual ar-
maments increased by a factor of 6. Paul Craig and Mark Levine similarly used
numbers of nuclear warheads by plotting the log of total warheads versus time.
They obtained a linear fit between 1968–85 as the number of warheads grew from
5000 to 25,000 strategic warheads.
P1: GFZ
Stable conditions exist between nations because of the “economic-burden” sec-

ond term. Finite labor, capital, and resources reduce production rates since a and
b “fatigue” terms are negative, damping the arms race. The gand h “grievance” or
“ambition” terms are positive when nations are dissatisfied or aggressive and are
negative when they are satisfied. A condition for stability results from solving the
first equation for Ny and substituting Ny and its derivative into the second equation
to obtain
d 2 Nx /dt 2 + (a + b)(d Nx /dt) + (a b − kl)Nx = hk + bg. (2.55)
The solutions can be analogous to those of the damped spring and mass system
in a gravitational field. Since a and b burden terms are positive, the dNx /dt term
provides damping. If the burden product ab is larger than the threat product kl, the
arms race is stable with damping. If ab is less than kl, arms race oscillations grow
and become unstable. If the two neighbors are happy (g < 0, h < 0), the solutions
are stable.
Problems
2.1 RV energy. A typical ballistic reentry vehicle (RV) rises to a height of 1000 km
above the Earth with a velocity of 7 km/s. If the RV has a mass of 400 kg, what
are its kinetic and potential energies at the top of its trajectory in kilotons?
2.2 Gravity. A rocket accelerates at a constant rate and constant angle 45◦ above
the flat Earth. It travels 200 s, obtaining a final velocity of 7 km/s. (a) What is
its final position? (b) What is the launch-weight to throw-weight ratio?
2.3 SS-18, three stages. The SS-18 is converted from two to three stages to save
fuel. Each stage adds one-third of the final velocity of 7 km/s. Assume mtwt =
8 tonne, Vex = 3.0 km/s, g = 0, and empty stage mass is 13% of initial mass.
What are the masses of the three stages?
2.4 Numerical trajectories. Assume a ballistic RV has a velocity of 7 km/s and it
is launched at 22◦ above the horizontal of a round Earth. (a) Using a Runge–
Kutta routine, how far along the Earth’s surface does the RV travel, how long
did the trip take and what was its peak altitude and acceleration? (b) How
does the range of the 22◦ launch compare to ranges for the launches at 15◦ and
30◦ ?
2.5 Mass concentrations. Using a simple numerical model, show that mass con-
centrations near launch sites affect a trajectory more when they are near the
target.
2.6 Gravitational bias. Approximate the ellipsoidal Earth as a sphere with an
extra 0.3% spherical mass and 0.15% negative masses at the poles. Using a
Runge–Kutta routine, determine the difference in trajectories for launches of
7 km/s along the equator and perpendicular to the equator.
2.7 GPS positions. (a) Show that unclassified GPS data could be improved by
referencing known positions on the trajectory using “differential GPS naviga-
tion.” (b) Show that millimeter accuracy could be obtained from two 1.5-GHz
signals.
P1: GFZ
Bibliography 53
2.8 GPS velocities. (a) How can GPS position data be used to obtain a velocity
vector? (b) An ICBM travels at 7 km/s directly under a GPS satellite traveling
4 km/s in the opposite direction. What is the Doppler shift of the 1.5-GHz
signal at 0, 1, and 10 s?
2.9 Superhardened silos. Assume a Russian missile has a single-shot kill proba-
bility of 0.9 and a reliability of 0.9. What are Pkill-1 and Pkill-2 for the silos that
are hardened by factors of 2, 5, and 10?
2.10 Big bombs. What are Pkill-1 and Pkill-2 if the yield is increased in problem 2.9
from 0.5 Mton to 1, 5, 50, and 100 Mton.
2.11 Dimensions. Obtain L = Y2/3 /CEP2 H 2/3 C by dimensional analysis.
2.12 Lethality limits. Show that very large yields or very good accuracy make the
lethality parameter irrelevant.
2.13 CEP. If the radial distribution of missile hits from an aim point p(r ) =
(1/2πσ 2 ) exp(−r 2 /2σ 2 ), show that integration of p(r ) to a radius CEP for an
SSKP = 0.5 gives CEP = 1.1 σ .
2.14 Smeared cookie cutter. The cookie cutter destruction probability can be
smoothed with a multiplicative function, similar to the Fermi-Dirac function,
g(r ) = 1/(1 + e x ). What is the form of x that makes logical and numerical
sense?
2.15 One-shot change rate. Derive the rate of change in kill probability for one
missile on a silo, Pkill-1 /Pkill-1 by varying CEP, H, R, and Y.
2.16 Two-shot change rate. Derive the rate of change in kill probability for two
missiles on a silo, Pkill-2 /Pkill-2 , by varying CEP, H, R, and Y.
2.17 Fratricide. A first warhead can destroy a silo with a probability R× SSKP.
Assume it cannot destroy the silo but still destroy a second warhead with re-
liability R. Use a fault tree to derive a fully effective fratricide, Pkill-2-fratricide =
2R × SSKP − R2 × SSKP. Compare this with Pkill-2 and show the range of pa-
rameters in which this is relevant.
2.18 Breakout from SORT. Design a force structure that is insensitive to attack by
many covert nuclear warheads and compare to the present US and Russian
forces.
2.19 Draw-down curve. Develop a graph that gives the number of surviving US
warheads under SORT as a function of the number of Russian warheads with
Pkill-1 = 0.7–0.9.
2.20 Conventional Richardson. Determine the stable and unstable solutions to
Richardson’s equation.
Bibliography
American Physical Society (1987). Science and Technology of Directed Energy Weapons, Rev.
Mod. Phys. 59, S1–S201.
——— (2003). Boost-Phase Intercept Systems for National Missile Defense, APS, College Park,
MD.
Blair, B. (1991). Strategic Command and Control, Brookings, Washington, DC.
Carter, A., J. Steinbruner and C. Zraket (1987). Managing Nuclear Operations, Brookings,
Washington, DC.
P1: GFZ
Herring, T. (1996). The global positioning system, Sci. Am. 274(2), 44–50.
Feiveson, H. (Ed.) (1999). The Nuclear Turning Point, Brookings, Washington, DC.
Hobson, A. (1989). ICBM vulnerability: Calculations, predictions and error bars, Am. J. Phys.
56, 829–836.
Hobson, A. (1991). The ICBM basing question, Sci. Global Secur. 2, 153–198.
International Institute for Strategic Studies. The Military Balance, Oxford Univ. Press, Oxford,
UK.
Levi, B., M. Sakitt, and A. Hobson (1989). The Future of the Land-Based Missile, American
Institute Physics Press, New York.
May, M., G. Bing, and J. Steinbruner (1988). Strategic arms after START, Int. Secur. 13, 90–113.
National Academy of Sciences (1997). The Future of US Nuclear Weapons Policy, National
Academy Press, Washington, DC.
——— (2002). Technical Issues Related to the Comprehensive Nuclear Test Ban Treaty, National
Office of Technology Assessment (1981). MX Missile Basing, OTA, Washington, DC.
Schields, J. and W. Potter (Eds.) (1997). Dismantling the Cold War, MIT Press, Cambridge, MA.
Schwartz, S. (1998). Atomic Audit, Brookings, Washington, DC.
Snyder, R. (1987). Approximations for the range of ballistic missiles, Am. J. Phys. 55, 432–437.
Stockholm Inter. Peace Research Institute. SIPRI Yearbook, Oxford Univ. Press, Oxford, UK.
Sutton, G. and D. Ross (1976). Rocket Propulsion Elements, Wiley, New York.
Wheelon, A. (1959). Free flight of a ballistic missile, Am. Rocket Soc. J. 29, 915–926.
P1: GFZ
3
The Defense: ABM/SDI/BMD/NMD
“. . . . to make nuclear weapons impotent and obsolete.”

[President Ronald Reagan, March 23, 1983]
3.1 ABM History

3.1.1 ABM
Attempts to develop defenses against strategic missiles began almost at the time
of their creation. Defenses have gone from antiballistic missile (ABM), to Strate-
gic Defense Initiative (SDI), to ballistic missile defense (BMD), to national missile
defense (NMD). The Soviets deployed the first ABM system, the Galosh, around
Moscow in 1962, which continues today in a modified version. The Johnson admin-
istration considered building the “Sentinel” defense to protect US cities, but this
would have been difficult because soft buildings extend over large urban areas. But
if cities could be completely defended, it would be possible for a nation to attack
first without fear of retaliation. It is clear that actual ABM systems would not be
able defend against a first strike, but if the nation that had a robust ABM system at-
tacked first, it might be able to defeat a weakened second strike. This is the famous
ABM strategic instability. Deployment of an ABM system can also be counterpro-
ductive since the existence of Galosh caused the United States to increase targeting
of Moscow. For these reasons, Johnson proposed the ABM and SALT treaties to
restrain both defensive and offensive weapons. Defensive constraints were origi-
nally rejected by Soviet leader Alexi Kosygin in 1968, as he stated that defensive
weapons were “moral.” Upon further thought he agreed to the ABM Treaty.
The Nixon administration negotiated the 1972 ABM Treaty, which limited each
side to 100 ABM launchers at two sites (later one site) and banned a nationwide
defense. This paved the way for the deployment of the “Safeguard” system to
defend US silos from Soviet attacks. Safeguard used Spartan missiles to attack
reentry vehicles (RVs) in their midcourse phase, 1000 km above the Earth. If Spartan
failed, then Sprint missiles would attack the RV in its reentry phase as it entered
the atmosphere. The exoatmospheric Spartan was similar to the Russian ABM-1B
55
P1: GFZ
56 3. The Defense: ABM/SDI/BMD/NMD
system and the high acceleration endoatmospheric Sprint was similar to the Russian
Gazelle system. Safeguard was deployed at a cost of $7 billion (in 1975 dollars),
but it was decommissioned after 6 months in 1976 because it was ineffective against
countermeasures and it was vulnerable to an attack against its radar.
3.1.2 SDI
In 1983, President Ronald Reagan made his “star wars” speech, stating that the
United States needed “a comprehensive and intensive effort to define a long-term
research and development program . . . on measures that are defensive . . . [to de-
stroy Soviet missiles] before they reach [US] soil or that of our allies.” The major
new concept of the SDI plan was to attack ICBMs in their boost phase before they re-
leased their RVs. Boost phase attacks on SS-18s could reduce the number of targets
by a factor of 10 as compared to attacks on the midcourse and reentry phases, but
such a strategy requires space-based weapons. The multiplier of 10 would be even
larger if SS-18s used countermeasures to protect the midcourse phase. The initial
SDI program was based on directed energy weapon systems (DEWSs) with space-
based lasers, particle beam weapons, rail guns, and other approaches. SDI was
shifted in 1987 from DEWS beam weapons to kinetic kill vehicle (KKV) weapons,
which would collide with RVs. One version, “brilliant pebbles,” was to remain in
orbit for a decade, always ready to accelerate toward an incoming RV.
3.1.3 BMD/NMD
Patriot missiles were ineffective against Iraq’s short-range Scuds in the first Gulf
War, partly because they were originally designed to attack planes and not mis-
siles. However, their use served a timely political purpose by calming Israel not
to respond to Iraq’s attacks. In 1993, the Clinton administration changed the SDI
program to the BMD program, which was intended to defend against theater mis-
siles with ranges up to 1000 km. A BMD weapon is the Theater High Altitude
Air Defense (THAAD) missile. Russia and China have been concerned that US de-
fenses might ultimately protect the entire United States, undercutting their “second
strike” deterrent. The coupling between defense and offense is scenario dependent,
making mixed offense-defense agreements difficult to negotiate.
Beginning in 2001, President George W. Bush gave strong support for protection
of the entire United States through an NMD with increased budgets. In his role as
president he had the power to withdraw the United States from the ABM Treaty,
which he formally did in June 2002. The United States is examining all technical
defense options, but it is somewhat vague about which systems might be deployed.
The Defense Science Board has considered a return to the nuclear-armed ABMs of
the 1970s, but this seems to have been rejected.
The Bush II administration abrogated the ABM Treaty for the stated reason that
some proposed tests would violate the treaty and because deployment of ground-
based interceptors (GBI) for NMD would violate the treaty’s Article I, “Each party
undertakes not to deploy ABM systems for a defense of the territory of its country.”
To attack the midcourse phase the Bush administration proposed to base GBIs on
P1: GFZ
3.2. Target Interactions 57
land and on sea; to attack the boost phase they proposed space-based lasers, space-
based KKVs, and the airborne laser (ABL). The 2003 American Physical Society
(APS) report on Boost Phase Intercept System goes into greater detail on this topic.
However, our chapter covers the basic ideas and we achieve independent and
similar results.
Defenses can be spoofed in several ways: infrared (IR) detectors can be voided
if attacking RVs are cooled; RVs can be placed in large aluminized balloons to
obscure their location from radar and IR detectors; RV-shaped balloons can be
used as decoys; and radar can be confused by releasing small wires, called chaff.
In addition, many small bomblets of anthrax could be released in the boost phase
to overwhelm the defense. It is conceivable that missiles could be attacked in the
boost phase before they released the bomblets, but the time window for realizing
this is very narrow. All offense and defense systems have their vulnerabilities.
The 2002 US withdrawal from the ABM treaty and avoidance of enhanced ver-
ification measures in the SORT treaty seem to have ended opportunities for addi-
tional strategic arms control. This chapter examines ABM, SDI, BMD, and NMD
systems that use lasers of many types (chemical, excimer, free-electron, and nuclear-
pumped) and basings (space, air, land, and sea), neutral particle beams (NPBs),
electromagnetic rail guns, and KKVs. Countermeasures that can overcome these
defenses are described. Excellent references are the 1987 APS study on DEWSs and
the 2003 study on boost phase interception.
3.2 Target Interactions

When missile bodies are in the boost phase, they contain rocket fuel, making them
soft targets that can be destroyed with an energy fluence of about fl = 1–20 kJ/cm2 .
On the other hand, RVs do not contain rocket fuel and they are made of heat
resistant, nonmetallic materials. The RVs are hard targets that need a higher fluence
fl = 100 kJ/cm2 to be destroyed. If sufficient energy is deposited at a surface,
rapid vaporization creates an ablative shock that can warp and crack the object. A
penetrating beam can damage semiconductors at lower energy levels, a fact that
convinced DoD to develop radiation-hard, amorphous semiconductors.
Mirrors have been used in solar furnaces to raise the solar flux by a factor of
104 to 1 kW/cm2 , heating objects to 3000◦ C (3273 K). A pedagogical ABM example
uses the solar flux in space of so = 1.37 kW/m2 to destroy a black, soft target. The
destruction time for solar flux so to destroy a soft target with a hardness fl = 1 kJ/cm2
is several hours, which is long enough to lose heat in space by thermal radiation.
Ignoring heat transfer, the time to destroy the target is
td = fl/so = (1000 J/cm2 )/(0.137 W/cm2 ) = 2 h. (3.1)

The reradiated flux from an object at 3000◦ C is much less than the incident flux,
σ T 4 = (5.7 × 10−12 W/cm2 )(3273 K)4 = 65 W/cm2 , (3.2)

where σ is the Stefan-Boltzman constant and T is temperature in Kelvin.
P1: GFZ
3.2.1 Time to Melt/Vaporize

If heating is quick, heat transfer can be ignored. The time to melt (tm ) through a
plate of finite thickness is determined from a heat balance:
(1 − Rrefl )i e tm = ρ H[c s (Tm − To ) + L m ], (3.3)
where Rrefl is reflectivity, L m is latent heat of melting at the melting point Tm ,

ρ is density, H is plate thickness, To is initial temperature, i e is the energy
flux, called the “irradiance,” and c s is the solid state specific heat. An irradi-
ance of i e = 10 kW/cm2 melts an aluminum plate of thickness H = 0.2 cm in
tm = 0.3 s with the following parameters: Rrefl = 0.8, ρ = 2.7 g/cm3 , c s = 0.90 J/g-K,
L m = 400 J/g, Tm = 660◦ C, To = 20◦ C. The fluence to melt a 2-mm aluminum
sheet is
flm = i e × tm = (10 kW/cm2 )(0.3 s) = 3 kJ/cm2 , (3.4)
consistent with soft-target values. The spreading of energy on a rotating booster

raises the needed destruction fluence by a factor of π to fl = 10 kJ/cm2 . Melt
time is critical, but the time to soften is more relevant since it can be de-
stroyed because it becomes a softer material. The 2003 APS study used a hard-
ness of 3.2 kJ/cm2 for liquid-fueled boosters and 24 kJ/cm2 for solid-fueled
boosters.
3.2.2 Vaporization Impulse

Vaporization time is longer than melting time because boiling points of materials
are considerably higher than melting temperatures and latent heats of vaporization
are an order of magnitude larger than latent heats of melting. The time to vaporize
the melted plate (tv ) is obtained from the heat balance,
(1 − Re )i o tv = ρ H[c l (Tv − Tm ) + L v ], (3.5)
where Tv is vaporization temperature, c l is the liquid specific heat, and L v is

the latent heat of vaporization. The same irradiance of i o = 10 kW/cm2 on the
aluminum plate is used with c l = 1.1 J/g-K, Tv = 2453◦ C, L v = 10, 732 J/g to
give a vaporization time tv = 3 s. This process requires a vaporization fluence of
(10 kW/cm2 )(3 s) = 30 kJ/cm2 , consistent with the hard target value of 100 kJ/cm2 ,
when the factor of π for rotation is included. Impulses delivered by x-rays are
the most efficient. Nuclear explosions deliver x-ray energy in about 1 μs, energy
that is absorbed near the surface, vaporizing it. Gaseous debris pushes the sur-
face, giving an ablative shock impulse in “taps” (cgs unit of 1 dyne-second/cm2
or a mixed unit of 10−5 Newton-s/cm2 ). Dropping a dime from 1-cm height
gives an impulse of 35 taps when stopped inelastically and 70 taps for an elastic
bounce.
P1: GFZ
3.3. Nuclear ABMs 59
3.3 Nuclear ABMs

3.3.1 Exoatmospheric ABM
KKVs can be spoofed with decoys or with cooled RVs that reduce the RV infrared
signature. They can also be fooled by RVs in large balloons, or by the use of small
bomblets of bioweapons in the boost phase. Conversely, the offense might want
to disguise a warhead to look like a decoy. Because of this weakness, there has
been a call to revitalize exoatmospheric ABMs using nuclear weapons to attack the
midcourse phase above the atmosphere. Such a weapon would circumvent many
countermeasures by attacking all weapons within its kill radius. This nuclear ABM
would be similar to Spartan, which was deployed in Grand Forks, North Dakota,
for 6 months in 1975. At altitudes of hundreds of kilometers, the lack of air raises
the fraction of x-rays to 75% of the yield. The fluence deposited on RVs from a
1-Mton warhead 0.5 km from the target above the atmosphere is
fl = 0.75(1 Mton)(4.2 × 1015 J/Mton)/(4π )(0.5 × 104 cm)2 = 100 kJ/cm2 , (3.6)
enough to destroy hard-target RVs. The kill range is greater than 0.5 km because
x-rays deliver an ablative shock impulse that takes less energy. To understand this
we compute the x-ray fluence needed to vaporize a thin surface of aluminum. If the
temperature of nuclear weapons is 20 million K, the energy of the most probable
photons from Wien’s law is
E = 4.97 kB T = 8 keV, (3.7)
where kB is Boltzman’s constant. The x-rays penetrate 0.0035 cm into aluminum,
heating a cross-sectional mass,
mass/cm2 = (0.0035 cm)(2.7 g/cm3 ) = 0.0095 g/cm2 . (3.8)

From a heat balance, this mass is melted with 0.01 kJ/cm2 and vaporized with 0.1
kJ/cm2 . This reduction in hardness increases the kill range.
3.3.2 Endoatmospheric
Low-altitude interceptors equipped with nuclear weapons were devised to attack
RVs in their reentry phase. The 1970s Sprint interceptors had large, 100-g acceler-
ations letting them rise to 10 km in seconds. Because some atmosphere exists at
this height, this location is called endoatmospheric. For example, if a 10-kton nuclear
warhead sent out 1.3 × 1023 neutrons/kton (Section 1.7), it would give a fluence at
1 km of
fln = (10 kton)(1.3 × 1023 n/kton)/(4π )(105 cm)2 = 1013 n/cm2 . (3.9)
At sea level, the fluence is reduced by factors as large as 100, abut at 15-km altitude
it is reduced by only a factor of 2. Fast neutrons can damage semiconductor circuits,
fissile materials, and high explosives.
P1: GFZ
3.4 Particle Beam Weapons

Particle beam weapons might be able to defend a target in near “real time,” since
these beams travel at almost the speed of light. However, the Earth’s magnetic field
deflects charged particles, hence NPBs would have to be used to obtain straight
trajectories. Neutral beams are formed by passing energetic hydrogen (H− ) ions
through a stripper gas, a process that removes a single electron from each ion to
form neutral hydrogen. NPBs must attack ICBMs and RVs above the atmosphere
to avoid degradation by the air. Therefore, under SDI it was proposed that NPB
accelerators be based in space. This would allow boost-phase missiles to be attacked
before the missiles released their RVs.
3.4.1 Beam Current

A 1-m diameter beam fixed on an ICBM at a distance of 1000 km must be tracked
with angular resolution of θ = 1 m/106 m = 1 μrad. One way for an NPB to
destroy a missile would be to heat the explosives on nuclear warheads to their
500◦ C ignition temperature. This could be done with an energy density of
Q/V = cρT/M = (25 J/mole-◦C)(0.8 g/cm3 )(500◦ C)/(50 g/mole) = 0.2 kJ/cm3 ,
(3.10)
where the high temperature specific heat c = 3Rgas = 25 J/mole-◦ C, ρs mass density,
and M is molecular weight (about 50). A 250-MeV beam deposits about 5% of its
energy in a 1-cm thickness, requiring a beam fluence fl = (0.2 kJ/cm3 )(20 cm) = 4
kJ/cm2 . The beam also raises aluminum temperature by
T = (Q/V)M/cρ = (200 J/cm3 )(27 g/mole)/(25 J/mole-◦ C)(2.7 g/cm3 ) = 80◦ C,
(3.11)
which causes internal stresses and misalignment. A smaller dose of 25 J/cm3 shifts
circuit element thresholds, while 1000 J/cm3 destroys electronic components.
A fluence of 4 kJ/cm2 requires an energy pulse
E = fl × area = (4 kJ/cm2 )(π × 104 cm2 /4) = 3 × 107 J. (3.12)
To reduce the required current, the beam could be trained on the ICBM for a
longer period, perhaps 10 s. Doing this would require continuous tracking, since a
1-m diameter RV with velocity 10,000 m/s moves out of a 1-m beam diameter in
t = 1 m/104 m/s = 100 μs. The pulse energy is E = VIt, where V is voltage, I is
current, and t is engagement time. For an engagement time of 10 s to obtain 0.3 ×
108 J, the beam current must be
I = E/Vt = (0.3 × 108 J)/(2.5 × 108 volts)(10 s) = 8 mA. (3.13)
If resolution is poorer by a factor of 10 at 10 μrad, beam currents have to be increased
by a factor of 100. The Los Alamos Meson Physics Facility produced 150 μA of H
ions at 1 GeV, about two orders of magnitude below the necessary current level.
P1: GFZ
3.4. Particle Beam Weapons 61
Building, deploying, and maintaining accurate and reliable NPB weapons would
be a formidable task.
3.4.2 Beam Power

The electrical energy needed to make a beam pulse at 15% efficiency is
E electrical = (0.3 × 108 J/0.15)(1 kWh/3.6 × 106 J) = 50 kWh. (3.14)
For an energy comparison we note that it takes 0.4 kg of coal to produce a

kilowatthour of electricity. Thus, a beam-pulse at 15% conversion would require
20 kg of coal. Peak electrical power for a 1-s burst is
Ppeak = (0.3 × 108 J)/(1 s) = 30 MWe . (3.15)
An angular resolution of 10 μrad would require peak power to be 3 GWe . In

addition, continuous power of 100–700 kWe would be needed for “housekeeping
chores.” These results are consistent with the APS study, which concluded that
1-GWe peak power might be needed for NPB weapons and electromagnetic rail
guns. Nuclear power is the only known way to achieve such power levels in orbit.
For this reason the Strategic Defense Initiative Organization (SDIO) planned the
100-kWe SP-100 space reactor with a thermal power of 2.5-MWt . The APS panel
concluded that a hundred or more space-based nuclear power plants would be
needed for SDI. By charging capacitors in space, peak power requirements could
be reduced, but not without causing complications. SDIO also considered burst
reactors capable of multi-GWe bursts. (See Section 4.6 for a discussion of monitoring
nuclear power in orbit.)
3.4.3 Burning a Hole in the Atmosphere

If space-basing fails, NPBs might be based on the ground. We estimate that the
energy to make a 10-km path through the atmosphere is 10% of the energy of
a perfect beam. Stability and degraded angular resolution are further issues. In
addition, there are beam losses in the beam hole. One-gigaelectronvolt particles
lose about 0.2 GeV/km in the air, preventing their passage through the entire
atmosphere. It is conceivable this problem could be solved with a very intense
laser that burned a hole in the atmosphere, so to speak, by partially evacuating a tube
of air. The reduced air density might contain the beam in the tube with less energy
loss. The energy needed to burn a hole by reducing air density will be estimated.
If air density in the hole is reduced by a factor of 10, energy loss would be reduced
to 0.02 GeV/km or 0.2 GeV in the full 10-km atmosphere (20% of a 1-GeV beam).
Temperature must be raised from 300 K to 3000 K to reduce air density by a factor of
10. In an atmospheric tube measuring 1 cm2 by 1 km, the air mass per kilometer is
(10−4 m2 )(1000 m/km)(1.3 kg/m3 ) = 0.13 kg/km. (3.16)

P1: GFZ
The energy required to heat the 1-km tube is
Q = mcT = (0.13 kg)(103 J/kg-K)(3000 K − 300 K) = 3.5 × 105 J. (3.17)
The energy for a 10-km long hole through the atmosphere is 4 × 106 J, which is 10%
of the energy for a perfectly aimed and sized beam of 3 × 107 J (Eq. 3.12). Additional
energy is lost to air forced out of the hole and by scattered radiation.
3.4.4 Other NPB Roles

It is unlikely that NPBs would be used to attack ICBMs, but they have been consid-
ered for other roles. NPBs in orbit could destroy satellites more easily than ICBMs
since satellites are soft targets in predictable orbits. However, there are easier ways
to attack satellites. NPBs might also be used to determine if objects in midcourse
phase are decoys or warheads. Since energetic protons can fission nuclear materi-
als in warheads, releasing fission gamma rays, a warhead could be distinguished
from a decoy after it had been hit with protons. SDIO planned to use this infor-
mation to enhance defensive targeting. The most sophisticated idea for targeting
warheads, called “adaptive preferential targeting,” would work in the following
way: Accurate and early tracking by radar and optical sensors would be used to
identify the silo, which an RV is targeting. If a single RV were targeting at a silo,
defense weapons would attack the RV. If two or more RVs were targeting a silo,
defense weapons would not attack these RVs. In this way the effectiveness of a lim-
ited defense would be maximized to save the largest number of silos. However, for
this approach to succeed, the defense must be able to accurately separate warheads
from decoys, and the RV trajectories must be known very accurately early in the
trajectory. Acquiring this information on a timely basis would be difficult.
3.5 Laser Weapons

Aircraft can be destroyed with a 25-kJ laser pulse acting from a short distance.
Because clouds and rain absorb laser beams, ground-based lasers are not effective
against ICBMs. This limits them to good weather. The atmosphere also broadens
laser beams, a drawback that, to some extent, could be overcome with adaptive
optics. Space-based lasers could avoid these problems. In addition, space-based
lasers might reduce the number of targets if MIRVed ICBMs were attacked in the
boost phase, before their RVs were released. The half-angle diffraction-broadened
resolution of a beam is θ = 1.2λ/D, where λ is wavelength and D is diameter of the
lens or mirror. Since there is no irradiance at the cut-off angle, the effective size of
a beam is smaller than 1.2λ/D; thus, we ignore the 1.2 factor. This gives the beam
radius on target,
r = Rλ/D, (3.18)
P1: GFZ
3.6. Orbital Chemical Lasers 63
where R is target distance. The area of the laser beam is

A = πr 2 = π R2 λ2 /D2 , (3.19)
giving energy fluence in Joules per square centimeter on target of
fl = Pt(1 − Rrefl )/A (3.20)
with power P in watts, engagement time t in seconds, and beam area A in square
centimeter. Our estimates for required laser power are lower bounds since we
ignore reflection by setting Rrefl = 0. Inserting area A gives
fl = Pt D2 /π R2 λ2 . (3.21)
Multiplying both sides of the equation by R2 /t gives two forms for the brightness
B, the energy flux into a solid angle in watts/steradian:
B = flR2 /t = PD2 /π λ2 . (3.22)

Note that the first version of B is described in terms of mission requirements of
needed fluence in a time interval on target at a distance R. The second version of B
is described in terms of the weapon parameters P, D, and λ. The energy/steradian
(J/sr) needed to destroy a booster with a fluence of 10 kJ/cm2 at range R of 1000
km is
Bt = flR2 = (104 J/cm2 )(108 cm)2 = 1020 J/sr. (3.23)
Hard targets need about 100 times more energy than this, raising the necessary
energy per solid angle to 1022 J/sr. Thus, SDI laser systems should be developed
with the following minimum capability:
1020 J/sr < Bt < 1022 J/sr. (3.24)
For an engagement time of t = 0.1 s, the minimum laser brightness should be in
the following range:
1021 W/sr < B < 1023 W/sr. (3.25)
3.6 Orbital Chemical Lasers

The radiation from CO2 lasers at λ = 10.6 μ gives considerable diffraction-
broadening, precluding its use as a space-based military laser. To reduce diffraction-
broadening, the ground-based, 2-MW Mid-Infrared Advanced Chemical Laser
(MIRACL) in New Mexico uses deuterium fluoride (DF) at 3.6–4.0 μ, which readily
passes through the atmosphere. The hydrogen fluoride (HF) version at 2.6–3.0 μ
is not useable for ground basing since it is absorbed by the atmosphere, but it is
preferred for space-basing since it suffers less diffraction broadening. An HF beam
with a 3-m diameter mirror (D) has a half-angle of
θ = λ/D = 2.8 × 10−6 m/3 m = 0.9 μrad, (3.26)
P1: GFZ
five times greater than the 2.4-m Hubble Space Telescope with visible light. The HF
beam diameter at a distance of 1000 km is
d = 2θ R = (2 × 0.93 μrad)(1000 km) = 1.9 m. (3.27)
The power needed to develop a brightness of 1020 W/sr to destroy a booster in
1 s is
P = Bπλ2 /D2 = π(1020 W/sr)(2.8 × 10−6 m)2 /(3 m)2 = 270 MW, (3.28)
which is two orders of magnitude larger than MIRACL’s 2 MW. This agrees with
the 1987 APS conclusion that “output powers at acceptable beam quality need
to be increased by at least two orders of magnitude for hydrogen-fluoride and
deuterium-fluoride lasers for use as an effective kill weapon in the boost phase.”
It would be difficult to make enhancements to lasers that would reduce the
power level needed to destroy boosters. Ten-meter diameter mirrors made from
segmented smaller mirrors in the Keck telescope could, in principle, reduce laser
power by one order of magnitude to
P = (270 MW)(3m/10 m)2 = 24 MW, (3.29)
but deployment of such a system in space would require tremendous effort.
3.6.1 Fuel in Orbit

HF beam energy needed to destroy a booster at 1000 km with a fluence of fl =
10 kJ/cm2 is
E = π (10 kJ/cm2 )(100 cm)2 = 3 × 108 J. (3.30)
The fuel mass needed to destroy a booster at η = 10% efficiency is
m = (dm/d E)(E/η) = (0.5 kg/MJ)(3 × 108 J)/(0.1) = 1.5 ton, (3.31)
where the energy density of HF fuel is dE/dm = 2 MJ/kg. The amount of fuel
needed in orbit to combat 2000 ballistic missiles of the former Soviet Union is very
scenario dependent. If all the lasers were in the correct location at the correct time
to attack 2000 Soviet ICBMs with two shots per booster, the lower-bound estimate
would be (2000 ICBM)(3 ton/ICBM) = 6000 ton of HF.
3.6.2 Laser Availability

Only a fraction of orbiting lasers are available for use at a time, since at any point in
time most are unavailable in another part of their orbit. The number of lasers needed
for defensive action depends on the location of targeted silos and submarines. The
fraction of time on duty to attack, the duty factor, is less than 25% since only one-
fourth of the satellites are in the former Soviet one-half hemisphere at a time. But
even this estimate is too optimistic since lasers have a limited range and silos are
based in clumps. The duty factor for one randomly placed laser with a 1000-km range
P1: GFZ
3.7. Earth-Based Lasers 65
to attack a particular silo would be

duty factor = π (1000 km)2 /4π (6400 km)2 = 0.5%. (3.32)
However, this is not the correct value for a space-based laser, which might have to
attack many targets. Since, for example, many silos are located along the path of
the Trans-Siberian Railway, a better estimate would be
duty factor = (2000 km)(10,000 km)/4π (6400 km)2 = 4%. (3.33)
But, this estimate is too low, since SLBMs at sea are also targets, and there are
other silos. On the basis of this general discussion it is reasonable to suggest a
duty factor of about 10%. Yet this could be too optimistic since SLBMs could be
gathered in a small region, saturating the defense at that location by monopolizing
available lasers and then allowing 100% transmission. If the duty factor were 10%,
the fuel required would be 10(6000 ton) = 60,000 ton of HF in orbit. To put this in
perspective, the payload of 308 SS-18s, at 8 ton each, is 2500 ton. This is 4% of the
orbital 60,000 ton of HF needed, a calculation that does not take into account the
mass of lasers and other equipment. Thus, the defense payload greatly exceeds the
offense payload. Since launch-weight is 20 times the throw-weight, total launch-
weight is over 1 million tons.
3.7 Earth-Based Lasers

3.7.1 Fighting Mirrors
Because of the difficulty of placing chemical lasers in orbit, SDIO explored the no-
tion of Earth-based lasers for attacking ICBMs. Lasers based on the ground cannot
attack boost-phase missiles because the curvature of the Earth blocks laser beams.
The only way for ground-based lasers to attack the boost phase is to use “relay
mirrors” to reflect beams down to “fighting mirrors” that focus the beams onto
boost-phase missiles. If the relay mirrors are located in geosynchronous orbit, only
a few would be needed because their relative position is fixed. However, these
relay mirrors would have to be very large to reflect the large beam size at that
distance. If the relay mirrors were located at lower altitudes of 5000 to 10,000 km,
the size of the relay mirrors could be smaller, but more mirrors would be needed
since they would not remain above Russia. The APS estimated that the energy from
large, land-based excimer lasers and free-electron lasers (FELs) was 3–4 orders of
magnitude too small to attack the ICBM boost phase over the former Soviet Union.
3.7.2 Adaptive Optics

Astronomers use adaptive optics (AO) to remove the broadening effects on stellar
images caused by atmospheric turbulence, thus improving resolution of land-based
telescopes to the diffraction limit. AO can also reduce effects of turbulence on
laser beams, going upward through the atmosphere from ground-based mirrors.
P1: GFZ
AO flexes the main mirror to geometrically shape laser pulses to overcome the
just-measured effects of atmospheric turbulence. Here is how it is done: Looking
upward, AO observes light stimulated by a second laser to measure turbulence
in the high atmosphere. Wave front sensors determine the slope of distortions in
the incoming wave front, which gives electronic information to deform the main
mirror by a submicron amount in order to modify the upward-traveling wave front.
Turbulent air cells in the troposphere are about 0.05 m in size at an altitude of 10 km,
which limits angular resolution of ground-based telescopes to
θ = cell size/altitude = 0.05 m/10 km = 5 μrad = 1 arc · s. (3.34)
This limit is five times larger than the diffraction limit of a 3-m mirror (0.5 μ/3 m =
0.2 μrad). Actuators deform a flexible mirror in less than 1 ms since the period of the
atmospheric disturbances is about 10 ms. Since AO must track moving relay mirrors
on satellites, it must take into account their motion. During 1 ms, a geosynchronous
satellite moves a distance of (3000 m/s)(1 ms) = 3 m, hence accurate tracking is vital.
3.7.3 Free-Electron Lasers

FELs have attained peak power of over a megawatt at wavelengths of 1 micron. The
laser action of a FEL results from the motion of a beam of electrons as it passes over a
periodic array of magnets, alternating in the transverse, up–down direction. When
relativistic electrons pass over an undulating magnetic field, the electrons perceive
a shorter magnetic periodicity in their reference frame. A Lorentz transformation
reduces the magnet array wavelength of centimeters to hundreds of microns in the
electron’s frame, according to
λ = λundulator /γ (3.35)
with γ = 1/(1 − v2 /c 2 )1/2 . In the laboratory frame, the electron experiences a trans-
verse oscillating force from the oscillating magnetic field. In the frame of the elec-
tron, the force on the electron from the electric field in the frame of the electron
of energy E = (γ v/c)B oscillates as the magnetic field B oscillates, causing dipole
radiation. The alternating field accelerates the electron from side to side with a
higher frequency because the magnets appear to be much more closely spaced. A
second Lorentz transformation contracts the wavelength in the laboratory frame
because the photons are emitted from moving electrons. The wavelength of the
radiation in the laboratory frame at the forward angle of θ = 0 is approximately
λ = λundulator /2γ 2 . (3.36)
The beam energy needed to obtain radiation at 1-μ wavelengths with a 4-cm
undulator is computed from the following equations:
γ 2 = λundulator /2λ = 0.04 m/2μ = 2 × 104 , (3.37)
giving γ = 140. Thus,
E beam = (γ − 1)mc2 = (139)(0.511 MeV) = 70 MeV. (3.38)
P1: GFZ
3.8. X-ray Laser Pumped with a Nuclear Explosion 67
More than 1 GW is needed to power FEL weapons because their Earth-based

laser beams would have to travel to geosynchronous orbit, as compared to space-
based lasers, with ranges of 1000 km that require only 20–200 MW. One cannot
simply use ratios of distances squared when configuring FEL weapons because
the scenario depends on the “battle-mirror” distance of closest approach to boost-
phase ICBMs. The 1987 APS study concluded that peak powers of 100–1000 GW
would be necessary for FELs, a power range that is far greater than the 1-MW
peak power achieved at 1 μ. In 1999, the Jefferson Laboratory obtained a continuous
FEL power of 1.7 kW at 3.1 μ. Another possibility is the excimer laser, which uses
krypton fluoride (λ = 0.25 μ), but its power level is too small by several orders of
magnitude.
3.8 X-ray Laser Pumped with a Nuclear Explosion

SDI plans were primarily for space-based weapons in low-Earth orbit, but Edward
Teller sought another approach because he realized that space-based weapons were
vulnerable to attack. He envisioned “pop-up missiles” based on submarines, which
could quickly leap into space to attack missiles in the boost phase. Teller and Lowell
Wood contemplated powerful x-ray lasers pumped with energy from a nuclear
weapon. The limited tests of these principles were not successful, a fact that was
kept from many policy makers. Other issues raised in this section show that even
the basing of a perfect system has very serious problems.
The curvature of the Earth requires basing submarines in the Arctic Ocean, closer
to the former Soviet ICBM silos. By being closer, pop-up missiles could be ready to
attack the boost phase in a more timely manner and with more energy on target.
This entails a race of time between boost-phase duration and time for the x-ray
laser (Excalibur) to rise sufficiently to attack the boost phase. If the response time
and Excalibur’s acceleration were good, the laser could look down on the rising
ICBM. However, if Soviets used fast-burn boosters to release RVs in one minute,
the pop-up laser would be too late. But there is a downside to being closer, since
close-in submarines would be more vulnerable to attack by the former Soviets.
3.8.1 Missile-to-Missile Distances

The most southerly Soviet missiles were located at 46◦ north latitude in Leninsk,
Kazakhstan. The closest part of the Arctic Ocean to Leninsk is in the Kara Sea at
71◦ N. The more distant locations of Franz Josef Land (81◦ N) and the North Pole
are safer but their locations greatly complicate the time constraints, and the extra
distance diminishes x-ray fluence on a target. Using locations of 30◦ and 45◦ north
of Leninsk, the arc length distances between submarine and silos are
S = Rθ = 6600 km(30◦ /57.3◦ ) = 3500 km (3.39)
from Franz Josef Land and 5200 km from the North Pole. These distances are much
greater than the 1000–3000 km kill radii discussed in the literature.
P1: GFZ
d
d H
H H 80 km
θ
θ
R R
Figure 3.1. Laser trajectories. For the case of a slow-burn booster, Excalibur rises to altitude
H (above the Earth’s radius R) to shoot a beam a distance d to a missile that is preparing to
launch RVs at altitude H. The minimum distance d is the case of equal altitudes of missile
and Excalibur, as shown on the left. For the case of a fast-burn booster, the RVs are released at
lower altitude. Excalibur must climb to higher altitudes and the beam must travel further, as
shown on the right. Atmospheric absorption of x-ray beams protects missiles up to altitudes
of about 80 km.
ICBMs are protected from x-rays by the atmosphere until they reach a height of
80 km. At this point the ICBM is 80 km beyond the radius of the Earth at about
6500 km above the center of the Earth. There are two ways to aim Excalibur: (1)
Excalibur rises to a height H above the Earth with the laser beam passing 80 km
above the Earth at θ/2, allowing the SS-18 to rise to the same height H. This is
advantageous since it takes less time for Excalibur to rise to a modest height, and
it provides more time since the SS-18 must rise to altitude H = d/(8R)1/2 , where
d is the distance between Excalibur and the missile and R is the radius of the
Earth. (2) Boost times for SS-18 and Peacekeeper are 5 and 3 min, respectively, but
these could be reduced to 1 min with the fast-burn booster. If the RVs are released
from a fast-burn booster at 80-km altitude, Excalibur must shoot from much higher
altitudes. The table shows that the beam distances would not be dramatically dif-
ferent, but the travel distances for Excalibur are four times further, requiring very
fast pop-up missiles. If the boost is terminated at an 80-km altitude, it is possible to
use the atmosphere as a protective cover to negate Excalibur. Ultimately, the time
window would be closed for Excalibur if fast-burn ICBMs were adopted (Fig. 3.1;
Table 3.1).
Table 3.1. Distance to target. The slow burn scenario has Excalibur
and SS-18s at the same height. The fast-burn scenario has the ICBM
release its RVs at 80 km, requiring Excalibur to rise higher. The
angular separations between target and Excalibur are 30◦ and 45◦ .
H (excaliber) H (ICBM) d to target
Slow-burn ICBM
θ = 30◦ 310 km 310 km 3500 km
θ = 45◦ 610 km 610 km 5400 km
Fast-burn ICBM
θ = 30◦ 1100 km 80 km 3700 km
θ = 45◦ 2800 km 80 km 7100 km
P1: GFZ
3.8. X-ray Laser Pumped with a Nuclear Explosion 69
3.8.2 X-Ray Laser Fluence

A distant missile would be destroyed if an x-ray beam could focus a significant
fraction of the yield of a nuclear weapon. In the 1980s, x-ray lasers were developed
by pumping with optical lasers. Very intense visible laser light is focused onto a
thin wire vaporizing it so quickly that it implodes. A selenium wire is used for
the wire, ionizing its atoms 24 times and lazing at λ = 20.63 nm and at 20.96 nm
(60 eV). In the 1990s, pulses from the Nova laser created tabletop x-ray lasers at
60 eV that lazed in a single pass since there were no credible x-ray mirrors for
this task. A highly focused horizontal line pulse (5 J, 0.1 ps, 150-μ wide) produced
instant plasma at 10 MK on a metal surface. The nickel or neon ions lazed, pro-
ducing output energies of 1 μJ for a conversion efficiency of 10−6 . The angular
spread of the beam was about 85 μrad, 15 times the diffraction limit. It has been
claimed that some x-ray lazing took place after a nuclear explosion test of Excal-
ibur, but the US General Accounting Office deemed the relevance of this effect was
overstated.
Without laser enhancement a 100-kton nuclear weapon produces an insignificant
fluence at 3000 km—by a factor of 107 too low to destroy a booster:
fl = (0.75 x-ray)(4.2 × 1014 J)/(4π )(3 × 108 cm)2 = 2.8 × 10−4 J/cm2 . (3.40)
Geometrical ray optics dictates that x-rays be emitted within a half angle of
θ G = D/L, where D is the rod diameter and L is the rod length (Fig. 3.2). This
gives better resolution from smaller diameter rods, but thinner rods produce wider
beam spreads because of diffraction-broadening, θ D = λ/D. If the two effects are
additive, the half-width is
θ = θD + θG = λ/D + D/L . (3.41)
The narrowest beam is obtained by taking the derivative of θ with respect to D,

which gives the rod diameter required to obtain the minimum spread,
Dmin = (λL)1/2 and θmin = 2(λ/L)1/2 . (3.42)
Figure 3.2. X-ray laser geometrical resolution. The fiber rod has diameter D and length L
for the single-pass laser.
P1: GFZ
If these were 1-keV x-rays (λ = 1.4 nm) and the rod was 1-m long, the rod diameter
would be small,
Dmin = (λL)1/2 = (0.14 × 10−8 m × 1 m)1/2 = 40 μ. (3.43)
The angular spread would be
θmin = 2(λ/L)1/2 = 2(1.4 nm/1 m)1/2 = 70 μrad. (3.44)
A 10-m rod reduces θ min to 25 μrad, which is almost 100 times larger than a
resolution of 0.3 μrad, to obtain a 1-m beam diameter at 3000 km. It has been
suggested that internal reflection in cone-shaped fibers can reduce the angular
spread, but Excalibur remains very speculative.
Tabletop x-ray lasers can convert input optical energy into x-rays with efficiency
η = 10−6 . Rather than use this value, we estimate the η needed to obtain lethal
fluence at 3000 km. The beam area at R = 3000 km is
A = πr 2 = π(θ R)2 = π(70 × 10−6 × 3 × 106 m)2 = π (210 m)2 = 1.4 × 105 m2 .
(3.45)
To obtain a fluence fl = 3 × 107 J/m2 from a yield Y = 30 kton weapon, efficiency
η must be
η = flA/Y = (3 × 107 )(1.4 × 105 )/(1.3 × 1014 J) = 0.03 = 3%, (3.46)
an estimate that is very optimistic. Obtaining the combined requirements for pop-
up basing with a high-efficiency Excalibur that can be aimed accurately is a tremen-
dous and difficult task, even without considering the possible defensive counter-
measures.
3.9 Kinetic Kill Vehicles

The SDIO shifted its emphasis from directed-energy beam weapons to space-based
KKV in 1987. The first KKV was the space-based interceptor (SBI), which was to be
based in groups of 10 in large satellite garages. The KKVs were expected to attack the
boost phase at a distance of 300 km from higher orbits in 50 s with relative velocity
6 km/s (300 km/50 s). Timing requirements can be relaxed if the midcourse phase
is to be attacked. After SBI stalled, “brilliant pebbles” were proposed to collide
with midcourse RVs. These were 1-m long, self-propelled satellites of 50 kg. They
were to be based in space, where each awaited a call to move toward boost-phase
and midcourse-phase missiles. Thousands of pebbles would be needed to cover all
such situations and they would have to function without maintenance for at least
a decade.
As SDI made very little progress toward deployment, changes in missile defense
were encouraged. Patriot missiles, which were designed to attack airplanes and not
missiles, were used to attack Iraqi Scuds in the 1991 Gulf War. The success of Patriot
was exaggerated and its overstated success reenergized the quest to develop means
of attacking theater missiles with ranges of 300–600 km and then to create an NMD.
Radar and IR sensors are used to give the trajectory to the defense missile, which
P1: GFZ
3.9. Kinetic Kill Vehicles 71
is then guided to the target for a kinetic kill or an explosion. The kinetic energy
density of KKV projectiles is
KE = mv2 /2 = (1 kg)(5000 m/s + 3000 m/s)2 /2 = 30 MJ/kg, (3.47)
where 5 km/s is the velocity of a theater missile and 3 km/s is the velocity of
THAAD. This energy density is much larger than that of explosives at 4 MJ/kg.
Even a 30-g pebble has a considerable energy of 1 MJ. The GBI is a kinetic kill
weapon that would be used to attack the ICBM midcourse phase with KKVs at
7–8 km/s. It is the first defensive weapon deployed as part of an NMD system in
Alaska and California to defend against a hypothetical missile attack by a dozen
missiles from Asia. As a follow on, these sites will have a total of 100 GBIs, with
another 125 GBIs being planned for deployment in North Dakota, but the pace is
slower than originally stated. The 2003 APS report analyzed sea-based intercep-
tors for attacking the boost phase. It concluded that “. . . existing US Navy Aegis
Standard Missile 2 could engage short- or medium-range ballistic missiles launched
from sea platforms without significant modification, provided that the Aegis ship is
within a few tens of kilometers of the launch platform.” In other words, the Aegis
must be deployed near known sea-launch platforms and it would not be effective
against sites on land.
The 2003 APS report analyzed the amount of mass needed in orbit for space-
based kinetic-kill vehicles: “. . . we find that a thousand or more inceptors would
be needed for a system having the lowest possible mass and providing a realistic
decision time. Even so, the total mass that would have to be orbited would require
at least five- to tenfold increase over current US space-launch rates, making such a
system impractical.
3.9.1 TMD and ABM Demarcation

While THAAD could be used to attack theater missiles, it could be upgraded to
have a strategic role. Early cueing data from radar and brilliant-eye sensors could
give THAAD extra time to increase the “footprint” it could defend. THAAD could
defend against strategic missiles if its velocity were increased to 5 km/s. The max-
imum closing velocity for theater intercepts is the sum of THAAD’s velocity of
3 km/s and the collinear theater missile velocity,
vTHAAD + vtheater BM, collinear = 3 km/s + 5 km/s = 8 km/s. (3.48)
The closing velocity against an ICBM is larger,
vTHAAD + vICBM, collinear = 3 km/s + 7 km/s = 10 km/s. (3.49)
The extra velocity of an ICBM raises closing velocity by 25%. However, the situ-
ation of ICBM and THAAD approaching at right angles,
vclosing = [(3 km/s)2 + (7 km/s)2 ]1/2 = 7.6 km/s, (3.50)
gives a closing velocity similar to the theater case. THAAD upgrades might create
a national defense, but the intent of GBI would be to fulfill that role.
P1: GFZ
The range of THAAD launched at 45◦ on a flat Earth is
R = vo2 sin 90◦ /g = (9 × 103 m/s)2 /(10 m/s2 ) = 900 km, (3.51)
which is about the size of the THAAD lethal footprint (if early tracking data are
available). The United States and Russia agreed to an ABM protocol in 1997 that
would separate strategic and theater defenses by limiting interceptors to 3 km/s.
Tests were to be limited to targets with velocities less than 5 km/s with ranges less
than 3500 km. The demise of the ABM Treaty in 2002 precluded such diplomatic
solutions.
3.10 Airborne Laser

The GBI is the first system for deployment for an NMD, but the ABL is being
developed to destroy the boost phase. The ABL hopes to use a 3-MW chemical
oxygen–iodine lasers (COIL) to be located on a fleet of seven Boeing 747s by 2011.
COIL has a short wavelength of 1.3 μ, but if this should fail a DF laser at 3.8 μ
might be used. ABL systems would employ a 1.5-m diameter mirror that would
give a diffraction-broadened COIL half-angle of θ = λ/d = 1.3 μ/1.5 m = 1 μrad.
The beam area at 300 km would be
A = π(θ R)2 = π (10−6 × 3 × 105 m)2 = 0.3 m2 . (3.52)
The fluence with 30% absorption is
fl = 0.3Pt/A = (0.3)(3 MW)(t)/(0.3 m2 ) = 3 kW t/cm2 , (3.53)
where beam time t is in seconds. This assumes that energy on target is propor-
tional to the laser power, which is not true since thermal blooming degrades energy
on target for higher energy fluences. Timing would be difficult since a Scud or fast-
burn missile takes 30 s to clear the clouds, leaving the ABL only about 30 s to target.
ICBMs would give more to target after observation. The 2003 APS report states that
2 min would be available to disable a solid-fueled ICBM after a firing solution is ob-
tained and 3 min would be available for liquid-fueled ICBMs. But it is unlikely that
the beam would be on target for more than 1 s. This gives COIL a fluence of
fl = (3 kW/cm2 )(1 s) = 3 kJ/cm2 , (3.54)

which is the hardness value used by the APS for a liquid-fueled booster. The higher
hardness of a solid-fueled booster of 24 KJ/cm2 requires a longer time. However, im-
pediments abound. Atmospheric turbulence over a 300-km horizontal path greatly
degrades the beam. In addition, energy absorption in the atmosphere causes ther-
mal blooming, which diverts laser beams. ABL uses adaptive optics to minimize
these effects (Section 3.7), but multiple turbulent air cells sized less than the beam
act like multiple lenses and create multiple beam spots. In addition, clouds can hide
rising missiles and can confound ABLs. Because of the short reaction time needed
to attack the boost phase, ABLs would have to be on station to be relevant. And
P1: GFZ
3.11. AntiSatellite Weapons 73
they are vulnerable. On October 4, 2001, a misfired Ukranian surface-to-air-missile

destroyed Siberian Airlines flight 1812 from a distance of 250 km.
The 2003 APS report concluded with the following: “Because solid-propellant
ICBMs are more heat-resistant, the Airborne Laser’s ground range against them
would be only about 300 kilometers, too short to defend against solid propellant
ICBMs from either Iran or North Korea . . . . Countermeasures against the ABL could
include applying ablative coatings or rotating the ICBM to reduce the amount of
heat the missiles absorbs, launching multiple missiles to overwhelm the ABL’s
capabilities or attacking the aircraft carrying the laser.”
3.11 AntiSatellite Weapons

Satellites are the mainstay of US national security as they are the “eyes and ears” to
watch for missile attacks and other military matters. (A drawback of their role is that
a nation could believe a massive attack was imminent if its satellites were attacked,
perhaps responding too quickly in time of crisis.) Satellites are also the source of
national technical means (NTM) of verification, monitoring weapons on foreign soil
and helping select locations for arms control inspections. The national security and
arms control applications of NTM are a stabilizing factor of global security when
used properly. It might be understandable for a country to destroy enemy satellites
in time of conflict, but to threaten satellites can cause an overreaction from the other
side. For these reasons, the US Congress imposed a ban on antisatellite (ASAT) tests
in the 1980s.
In the 1980s, the United States deployed an ASAT missile on F15 fighters that
could lift KKVs into orbit, while the Soviets deployed ASAT missiles that used
fragmentation explosives. Ground-based lasers can destroy satellites that are soft
objects in predicable orbits, as happened when the MIRACL laser at White Sands,
New Mexico, attacked a retired US satellite in 1997. DF gas was used because its
3.8-μ radiation easily passes through the atmosphere. Assuming adaptive optics
constrains turbulence-broadening to twice the diffraction limit gives a half-angle of
θ = 2λ/d = 2(3.8 μ/3 m) = 2.5 μrad, (3.55)
for a 3-m diameter mirror. The beam area at 500 km is
A = π (Rθ )2 = π(5 × 105 m × 2.5 × 10−6 )2 = 4.9 m2 . (3.56)
Assuming 80% of the radiation is absorbed by the satellite, the absorbed fluence is
fl = 0.8(Pt/A) = 0.8(2 MWt/4.9 m2 ) = (0.03 kW/cm2 )t. (3.57)
If the ASAT beam tracks a satellite through a 60◦ arc above the ground-based
ASAT, it covers about 200 km. This gives an engagement time t = 30 s, giving an
absorbed fluence of about 1 kJ/cm2 that could destroy the satellite.
P1: GFZ
3.12 Rail Guns

SDI scientists developed electromagnetic rail guns whose function is to fire projec-
tiles in space. The large Checkmate rail gun projected 100-g pellets to 4 km/s in
1 ms with an acceleration
a = v/t = (4000 m/s)/(1 ms) = 4 × 106 m/s2 (3.58)
from a force
F = ma = (0.1 kg)(4 × 106 m/s2 ) = 4 × 105 N. (3.59)
A rail gun does not use a solid conducting armature to produce a magnetic force,
but rather it uses a plasma-arc armature heated to 10,000 K. The magnetic field
is produced between two linear conductors with the current going up one side,
crossing over through the plasma and returning down the other side. Checkmate
produced 200 kA at 40 kV from discharging stored energy in six capacitors. A 4-cm
spacing between the wire gives a magnetic field Bo at the 2-cm midpoint,
Bo = (2 × 0.5)(μo /4π)(2I /r ) = (10−7 T-m/A)(2 × 2 × 105 A)/(0.02 m) = 2 T. (3.60)
This is the minimum magnetic field at the center of the gap; it increases dramatically
near the two wires. We arbitrarily increase the average effective field over the
minimum field by a factor of 25 to obtain the force required by Eq. 3.59,
F = IL × 25Bo = (2 × 105 A)(0.04 m)(50 T) = 4 × 105 N. (3.61)
The pellet’s kinetic energy,
KE = mv2 /2 = (0.1 kg)(4000 m/s)2 /2 = 0.8 MJ, (3.62)
was obtained from discharging six 1-MJ capacitors producing 200 kA at 40 kV for
about 1 ms, giving a system efficiency of 15%. The power needs for deploying rail
guns in space would have to be met with nuclear power supplied in orbit.
Problems
3.1 Boost phase sine qua non. (a) What are two reasons why the boost phase of an
incoming missile is the best place to defend against an attack? (b) What kind
of basing is required for a defense that attacks the boost phase?
3.2 Layered defense. A hypothetical defense could consist of three layers, one
to attack the boost phase, one to attack the midcourse, and one to attack the
reentry phase. (a) If each layer has a kill probability of 90%, how many of
10,000 warheads will penetrate the defense? (b) If one layer is defeated by
countermeasures, how many warheads will penetrate the defense? (c) How
could the offense “punch a hole” in the defense?
3.3 Countermeasures. What countermeasures by the offense could defeat attacks
by the defense on the boost phase, midcourse phase, and reentry phase? De-
scribe the physics involved.
P1: GFZ
Problems 75
3.4 Solar ABM. If a solar ABM concentrated solar flux of 1.35 kW/m2 by factors
of 10–100, how long would it take to destroy soft and hard targets with 1–100
kJ/cm2 . Assume reflectivity rrefl = 0.5.
3.5 Rotating targets. If beam pulse is longer than rotational period, how much is
the fluence on target reduced?
3.6 Sprint and fast-burn. (a) How long would it take the Sprint ABM to rise 20 km
at 50 g’s. What would be its final velocity? (b) Fast-burn boosters can unload
RVs in 1 min at a height of 100 km. What acceleration would be needed for
this to occur, and what would be the final velocity of the booster?
3.7 Target hardness. A hypothetical RV is covered with 1 cm of black ice (ε = 1)
at 0◦ C. (a) What fluence will melt and vaporize this RV cover? (b) How long
would it take sunlight (1.35 kW/m2 ) to melt and vaporize the ice cover?
3.8 X-ray impulse loading. (a) Show that 8 keV x-rays can melt the penetration
depth of aluminum with a fluence of 10 J/cm2 and vaporize it with 120 J/cm2 .
(b) How many taps/cm2 are delivered if the penetration depth of aluminum
is vaporized at 2500◦ C?
3.9 NPB midcourse discrimination. (a) Devise an approach that uses NPBs to
determine if an object is a decoy or a nuclear warhead. (b) Devise a generalized
equation that projects counting rates at a detector as the result of an NPB with
a beam width of 1 μrad and an NPB H atom current of 1012 /s.
3.10 Beam width. The Rayleigh diffraction criterion for zero intensity is 1.22λ/D.
Using the diffraction intensity of (sin2 θ)/θ 2 , what is the beam half-width
needed for intensity to drop by 50% from its peak? How does this compare to
our approximation of λ/D?
3.11 Geosynchronous basing. (a) How many J/sr are needed to destroy soft and
hard targets (1–100 kJ/cm2 ) from geosynchronous orbit (GEO)? (b) If the en-
gagement time is 0.1–10 s, what is the required brightness in W/sr?
3.12 Orbiting lasers attack satellites. A 1-MW chemical laser using HF and a 2-m
diameter mirror attacks soft satellites 3000-km away. (a) How long an engage-
ment time is needed to destroy soft targets (1 kJ/cm2 )? (b) How much energy
is expended? (c) How much HF (2 MJ/kg) is burned at 15% efficiency?
3.13 GEO battle mirrors. A land-based 10-MW HF laser beam from a 3-m diameter
mirror is reflected off a relay mirror at GEO orbit. (a) What is beam size and
fluence at GEO? (b) Design the relay mirror needed to obtain a 3-m diameter
spot for the battle mirror in a low-earth orbit (LEO) with an altitude of 150 km.
3.14 Distance to target. (a) Derive equations to determine the distance from Excal-
ibur to target using the conditions of objects at equal height H and a fast-burn
booster rising 80 km. (b) Confirm the numbers in Table 3.1.
3.15 Stabilizing ABM. (a) In what ways would ABM systems be stabilizing and
destabilizing for the nuclear weapons states? (b) Use a numerical model to
prove your point.
3.16 ABM Treaty. What actions did the former ABM Treaty forbid and what did it
allow?
3.17 Radar power limit. The ABM Treaty limits the product of power and radar
area of outward looking radar at national boundaries to 3 MW-m2 . (a) What is
P1: GFZ
the radar irradiance on a 1 m2 RV, 5000 km from the radar, using θ = λ/D for
the 10-GHz, 30-m diameter Cobra Dane? (b) What is the reflected irradiance
at Cobra Dane from the RV with a reflectivity of 0.2? (c) What is the ratio of
reflected power to output power at Cobra Dane?
3.18 Adaptive preferential targeting. To maximize effectiveness how would you
deploy 10 defensive missiles with an effectiveness of 50% against 5 incom-
ing RVs, 2 aimed at a Peacekeeper with 10 warheads each and 3 aimed at
Minuteman III with 1 warhead each?
3.19 Boost-phase timeline. To protect against biological bomblets, high-
acceleration missiles carrying KKVs are based on ships in the Sea of Japan
to attack missiles from North Korea. (a) How long does it take for defensive
missiles to rise to a height of 100 km with a velocity of 7 km/s? (b) How long
does it take for the KKV to approach a missile 300-km away? (c) If the missile
with bomblets takes 1–3 min to gain a 100-km altitude, how much time does
the defense have to make an attack decision?
Bibliography
Aftergood, S., et al. (1989). Space arms control, Sci. Global Secur. 1, 55–146.
American Physical Society (1987). Science and Technology of Directed Energy Weapons, Rev.
Mod. Phys. 59, S1–S202 and Physics Today 40(3), S1–S16.
———(2003). Boost-Phase Intercept Systems for National Missile Defense, APS, College Park,
MD.
Carter, A. and D. Schwartz (Eds.) (1984). Ballistic Missile Defense, Brookings, Washington,
DC.
Fitzgerald, F. (2000). Way Out in the Blue, Simon and Schuster, New York.
Forden, G. (1999). The airborne laser, IEEE Spectrum 36(3), 40–49.
Garwin, R. (1985). How many orbiting lasers for boost-phase intercept, Nature 315, 286–290.
Gronlund, L., et al. (2000). The continuing debate on national missile defense, Phys. Today
53(12), 36–43.
Hey, N. (2006). The star wars enigma: Behind the scenes of the cold war race for missile
defense, Potomac Books, Washington, DC.
Office of Technology Assessment (1987). SDI Technology, Survivability and Software, OTA,
Washington, DC.
———(1985). Ballistic Missile Defense Technologies, OTA, Washington, DC.
Sessler, A., et al. (2000). Countermeasures, Union of Concerned Scientists, Cambridge, MA.
Taylor, T. (1987). Third-generation nuclear weapons, Sci. Am. 256(4), 30–38.
P1: GFZ
4
Verification and Arms Control Treaties
“Trust, but verify.”

[President Ronal Reagan to General-Secretary Mikhail Gorbachev, Dec. 8, 1987]
The two leaders just signed the 1987 Intermediate-Range Nuclear Forces (INF)
Treaty and were in the process of agreeing on aspects of the Strategic Arms
Reduction Treaty (START). The signing marked the beginning of the end of the
cold war, 2 years before the November 1989 fall of the Berlin Wall and 3 years
before the agreement on the Conventional Arms in Europe Treaty.
4.1 Verification Context

This chapter is a physics primer on verification technologies, which will be dis-
cussed within the context of arms control treaties. Nations remain in arms control
treaties because they judge the disadvantage of controls on nuclear weapons to be
less dangerous than a system with no controls. The failure of the United States to
ratify the Comprehensive Test Ban Treaty (CTBT) and its subsequent withdrawal
from the Anti-Ballistic Missile (ABM) Treaty are exceptions to a global consensus;
the world awaits what is in store for arms control.
It is important to distinguish between “monitoring” and “verification.” Monitor-
ing is data collecting through satellites and other means. The information collected
includes photographic, infrared, neutron, x-ray, gamma ray, electromagnetic pulse,
radar, seismic, hydroacoustic, radionuclide, infrasound, and communications data.
Optical resolution of details by US satellites improved from 12 m in 1960 to 0.7 m in
1980 to less than 0.1 m at present; the United States now sells photos with resolution
of 0.5 m. The KH series began with images on film that were jettisoned to Earth
by parachute. The advent of charge-coupled devices now allows direct electronic
transmission of digital images. By 1988 satellites began to transmit radar images. In
the atmosphere the Air Force uses RC–135 planes, the Navy uses EP-3 planes, and
the Army uses Predator drones to gather further information. These technologies
are an interesting scientific tours-de-force as well as being politically important. To
avoid being labeled “secret,” the information is described as obtained by National
77
P1: GFZ
78 4. Verification and Arms Control Treaties
Technical Means (NTM) of verification. NTM is buttressed with on-site inspec-

tions (OSIs) that take data mostly at declared sites, but occasionally at undeclared
sites. Interagency groups analyze the data and come to preliminary judgments to
determine if decision makers should be informed about compliance problems.
Verification is the quasi-judicial act that determines if the monitored data are suf-
ficient to prove that a treaty party has or has not violated the terms of an agreement.
The level of confidence in the data determines the adjectives applied to a violation,
such as possible violation, likely violation, or probable violation, or just plain viola-
tion. The charge of likely violation is close to a preponderance-of-the-data standard but
less than a standard of beyond all reasonable doubt. The verification process has been
politicized, but it has been mostly accurate. The Soviet Krasnoyarsk radar was an
unambiguous violation of the ABM Treaty. An example that cuts the other way was
the falseness of the US charge of a likely violation of the Threshold Test Ban Treaty
(TTBT) by the Soviets in the 1980s, an issue with which the author was intimately
involved as the State Department technical lead on testing issues.
4.1.1 Effective Verification

During the ratification of the INF Treaty, Ambassador Paul Nitze defined effective
verification: “ . . . . if the other side moves beyond the limits of the treaty in any mili-
tarily significant way, we would be able to detect such violation in time to respond
effectively and thereby deny the other side the benefit of the violation.” Thus, any
cheating must be detected in a timely manner before it can threaten national secu-
rity. Since nations are already threatened with a plethora of legal nuclear weapons,
the “effective” criterion concerns the degree of marginal threat due to cheating
beyond the legal level of threat agreed to by the treaty. During the ratification of
START, Secretary of State James Baker repeated this definition, but added a new
criterion: “Additionally, the verification regime should enable us to detect patterns
of marginal violations that do not present immediate risk to the US security.” Dur-
ing the negotiations the United States and Russia decided how much verification
was enough: there are wise verification measures and their are superfluous ones.
An excellent one from START is the reentry vehicle on-site inspection (RVOSI) that
counts RVs (warheads) on a missile. RVOSIs would be an important tool if limits
at lower numbers of warheads can be achieved.
4.2 Arms Control Treaties

The US executive branch selects policy options in interagency groups (IG) led by the
National Security Council (NSC), working with the State Department, the Defense
Department, the Joint Chiefs of Staff (JCS), the Arms Control and Disarmament
Agency (ACDA, 1961–99), the Central Intelligence Agency (CIA), and in some
cases the National Nuclear Security Agency (NNSA). If a dispute exists, the issue
goes to the NSC, led by the president and composed of the secretaries of State and
Defense, the chairman of the JCS and the director of the CIA. Finally, the president
P1: GFZ
4.2. Arms Control Treaties 79
makes the decision to endorse, amend, or reject the IG agreement. In this section
we list the central provisions of key arms control treaties that have passed through
this process.
4.2.1 Nuclear Testing Treaties

LTBT. The Limited Test Ban Treaty, entered into force (EIF) in 1963, bans nuclear
tests in the atmosphere, outer space, and under water.
TTBT. The Threshold Test Ban Treaty bans underground nuclear tests of over
150 ktons. Its 1988 protocol added OSIs. (Signed 1974, EIF 1990.)
PNET. The Peaceful Nuclear Explosions Treaty limits PNEs to underground explo-
sions to a maximum of 150 kton for individual PNEs and 1500 kton for group
explosions. (Signed 1976, EIF 1990.)
CTBT. The Comprehensive Nuclear Test Ban Treaty bans nuclear tests of any yield
at any place and any time. The CTBT created the International Monitoring System
(IMS) of 300 monitoring stations that use seismic, hydroacoustic, radionuclide,
or infrasound sensors. (Signed in 1996. As of July 2006 176 nations signed and
134 ratified. EIF will occur after 44 nuclear-power states have ratified, 34 of these
ratified. Three nuclear weapon states (NWSs) have ratified (France, Russia, and
United Kingdom) and two have not (China and United States). Ratification failed
in the US Senate by 51 to 48 in 1999, primarily over the issue of verification.
4.2.2 Strategic Nuclear Arms Treaties

SALT I. The Strategic Arms Limitation Talks (1972) led to a 5-year agreement that
passed Congress with a majority vote. SALT I capped the number of allowed
intercontinental ballistic missiles (ICBMs) and submarine-launched ballistic mis-
siles (SLBMs) at their 1972 levels, but it did not cover bombers.
ABM Treaty. The Anti-ballistic Missile Treaty (1972) constrained each side (United
States and Union of Soviet Socialist Republics) to 100 ABM launchers located
at no more than one site (1974). The ABM Treaty forbade national area defense
and radar construction within a nation and it created the Standing Consultative
Commission (SCC) to implement the SALT and ABM treaties. SDI beam weapons
were constrained by a provision that certain other physical principles (such as
beam weapons) “would be subject to discussion.” President George W. Bush
withdrew the United States from the ABM treaty effective June 2002.
SALT II. The unratified SALT II limits (1979–1986):
r 2400 ICBMs, SLBMs, and heavy bombers
r 1320 MIRVed ICBMs, SLBMs, and heavy bombers with cruise missiles
r 1200 MIRVed ICBMs and SLBMs
r 820 MIRVed ICBMs
r 1 new ICBM type during the 5-year duration.
INF. The Intermediate-Range Nuclear Forces Treaty (1987, EIF 1988) bans all ballis-
tic and cruise missile systems everywhere between ranges of 500 km and 5500 km
and established notifications and OSIs of historic proportions. By 1991 the United
P1: GFZ
States eliminated 283 launchers and 859 missiles capable of carrying 859 war-
heads, and the Soviets eliminated 825 launchers and 1846 missiles capable of
carrying 3154 warheads.
START I. The first Strategic Arms Reduction Treaty (1991, EIF 1994) set the following
limits:
r 1600 ICBMs, SLBMs, and heavy bombers
r 6000 accountable warheads (about 8000 actual warheads)
r 3600 tons throw-weight (46% reduction, mostly from cutting 50% of 308 SS-18s)
r 13 types of OSIs, and perimeter portals at mobile-missile production facilities.
START II. Signed in 1993, ratified by US Senate in 1996 and by the Russian Duma
in 2000, but it did not enter into force because of US withdrawal from the ABM
treaty. It set the following limits:
r 3000–3500 warheads on ICBMs, SLBMs, and heavy bombers (fully counted)
r elimination of all heavy SS-18 missiles
r elimination of MIRVed ICBMs except for
– 90 SS-18 silos converted for SS-25/27
– 105 SS-19s, which must be reduced from 6 warheads to 1.
SORT. The START III framework was agreed to in 1997 by Presidents Clinton and
Yeltsin. It called for a limit of 2000–2500 warheads and enhanced transparency
on warhead inventories and destruction. The Strategic Offense Reduction Treaty
(SORT) was signed in May 2002 by Presidents Bush and Putin, and ratified on
March 6, 2003. It limits operational warheads to 1700–2200 by 2012. This is the
START III limit as it does not count 240 warheads on 2 Trident submarines in
overhaul. SORT expires on the day the limits come into affect and it does not
include intermediate compliance dates. SORT does not include additional verifi-
cation measures. In 2012 the United States will have 2200 operational warheads,
plus 240 warheads on 2 Tridents in overhaul, plus an active reserve, responsive
force of 2000 warheads, which could be deployed in “weeks, months or years”
depending on the delivery platform. In addition, the United States will have
additional warheads in the inactive reserve stockpile, plus a stockpile of primary
and secondary stages. Russia did not mind the demise of START-II since they
wanted to keep 138 SS-18s and many SS-19s.
4.2.3 Other Treaties and Agreements

BWC. The Biological Weapons Convention (1975) bans production or acquisition
of biological agents for weapons, but does not establish monitoring.
NPT. The Nuclear Non-Proliferation Treaty (1970) prevents the spread of nu-
clear weapons, while promoting the peaceful uses of nuclear energy. Creates
two classes of nations, the Nuclear Weapon States (NWS: the United States,
Russia, United Kingdom, France, and China) and 180 nonnuclear weapons states
(NNWS), with the exception of India, Israel, and Pakistan. The NPT was extended
without a time limit in 1996. The NPT key agreements are as follows:
r NWS not to transfer/assist nuclear weapon technology to NNWS.
r NNWS not to receive/acquire nuclear weapons technology.
P1: GFZ
4.3. Optical Reconnaissance 81
r NNWS accept International Atomic Energy Agency inspections on nuclear

facilities.
r NWS to assist NNWS with their peaceful nuclear power programs.
r NWS volunteer for a few inspections.
r NWS undertake negotiations on nuclear disarmament.
MTCR. The Missile Technology Control Regime (1987) bans export of ballistic mis-
siles with a range of over 300 km. In 2002, the International Code of Conduct
Against Ballistic Missile Proliferation was signed by 40 states, but not China,
India, Iraq, Iran, North Korea, and Pakistan.
CWC. The Chemical Weapons Convention (1996) bans production, acquisition,
stockpiling, transfer and use of chemical weapons with extensive verification.
Nuclear Weapons-Free Zones. Antarctica (1961), seabeds (1972), outer space and moon
(1967), South America (Treaty of Tlatelolco, partial EIF in 1968, more complete
with 23 nations by 1989), South Pacific (Treaty of Rarotonga, signed 1996), Africa
(Treaty of Pelindaba, signed 1996), and Central Asia (6 former Soviet Republics,
signed 2002, but not EIF).
CFE. Conventional Armed Forces in Europe Treaty (1992) ended cold war in Europe,
allotting equal forces to NATO and Warsaw Pact. The Warsaw Treaty Organiza-
tion reduced from 200,000 to 80,000 tanks, artillery, personnel carriers, aircraft
and helicopters, while NATO had essentially no reductions.
Hot Line. Hot line agreements (1963, 1984) assure quick and reliable communica-
tions between the United States and Russia. The Nuclear Reduction Risk Centers
Agreement (1987) serves as a cool line to transmit large amounts of arms control
information between Washington and Moscow.
Open Skies. Commits 27 Eurasian and North American nations to open their airspace
to overflights for photography and radar reconnaissance. (Provisionally applied
in 1992, EIF 2002).
Global Organizations. United Nations Security Council and General Assembly
(1945), Zangger Committee (nuclear trigger list, 1971), Nuclear Suppliers Group
(export criteria, 1975), United Nations Conference on Disarmament (1979),
Australia Group (chemical weapons, 1985), Wassenaar Arrangement (conven-
tional arms technologies, 1996), Ottawa Group (1996, land mines).
4.3 Optical Reconnaissance

Reconnaissance optical systems have improved considerably in many ways:
r Aerial survey films are available with high contrast of 1000 lines/mm (1 μ between
lines).
r The advent of charge-coupled devices (CCD) in the 1980s allow real-time pho-
tography at good resolution.
r Chromatic aberration is reduced with computer-aided lens design and machin-
ing.
r Blurring by satellite motion is removed by moving camera with precision servo-
controls.
P1: GFZ
r Similar-appearing objects can reflect and emit differently in the visible and IR,
measured with multilens, multispectral cameras at many wavelengths.
r Two views of the same terrain at different angles gives stereoscopic images to
obtain heights. of objects. Cartographic cameras cover 25,000 km2 in stereo with
a 10-m resolution, using only two pictures.
r Fourier transform infrared spectroscopy rapidly and accurately determines trace
impurities in air.
4.3.1 Resolution of Film Versus CCD

The half-angle resolution of the Hubble Space Telescope with a 2.4-m diameter (D)
mirror is
θ = 1.2λ/D = 1.2(0.5 μ/2.4 m) = 0.25 μrad. (4.1)
Film and CCD resolution is limited by size of film grains and CCD pixels. The
thin lens equation for satellites is simplified because distances to objects (o) are
much larger than lens focal length ( f ) of lens (o = 150 km f ). Thus, the image
distance (i) is equal to the focal length, from
1/ f = 1/i + 1/o ≈ 1/i or i = f. (4.2)
CCD resolution is limited by a pixel size of about 10 μ, which is about the same
for film of 100 lines/mm (1 line/105 m). Better film with 1000 lines/mm gives a
resolution of 1 μ. The object size from an image size of h i = 10 μ (CCD or lesser
film) and a 6-m focal length (i = f ) is
h o = h i (o/i) = (10−5 m)(12 × 104 m/6 m) = 20 cm. (4.3)
Diffraction broadening of 0.25 μrad is ignored since the geometrical spreading from
10-μ film/pixel is much bigger,
θ = h i /i = 10−5 m/6 m = 1.7 μrad. (4.4)
If better film of 1000 lines/mm is used, geometrical resolution is 0.17 μrad, which
must be combined with diffraction broadening to give
θ = [(0.17 μrad)2 + (0.25 μrad)2 ] = 0.3 μrad. (4.5)
Object resolution with good film at an altitude of 150 km can be
h o = θ(150 km) = (0.3 × 10−6 )(15 × 104 m) = 5 cm. (4.6)
The Hubble Space Telescope uses a CCD array of 1600 × 1600 pixels, each pixel
measuring about 10 μ, to obtain resolution of 0.4 μrad. With 5-cm resolution, it
would not be possible to read Pravda’s masthead from orbit, but it would be possible
to distinguish a VW Beetle from a Saturn. CCDs have become the verification sensor
of choice with a small resolution sacrifice because of the advantages listed below:
r CCDs are used in real time with computers, while film requires time delays and
scanning.
P1: GFZ
4.4. Adaptive Optics 83
r CCDs are reusable, removing considerable weight in orbit.

r CCDs are sensitive to a broader range of wavelengths.
r CCDs have a much greater black/white dynamic range.
r CCD response is extremely linear compared to film’s great nonlinearity.
r CCDs are 70 times more light-efficient than film.
4.4 Adaptive Optics

Stars twinkle at night because atmospheric turbulence continually changes refrac-
tion angles. Luckily for reconnaissance photography, angular, “seeing” resolution
for looking downward from a satellite is considerably smaller than for the case
of looking upward from a telescope, which is about 5 μrad (1 arcsecond). Seeing
from satellites is better than seeing from ground-based telescopes because of a re-
versal of the relevant distances. For telescopes, the object-to-turbulence distance
is much greater than the turbulence-to-lens distance of 1–10 km. For satellites the
opposite is true; the object to turbulence distance of 1–10 km is much less than
the turbulence-to-lens distance of 100 km. It is this difference that makes “seeing”
downward better than “seeing” upward. This can be proven experimentally by us-
ing a 0.5-m converging lens to establish an image of a wire-mesh that is about 5 m
from the lens. First, observe the object when a Bunsen burner is placed near the ob-
ject. Second, observe the object when the Bunsen burner is near the lens. The burner
turbulence does not noticeably distort the image when it is close to the object, but
it does distort the image when it is placed near the lens. A second experiment:
When a person is taking a shower with a translucent shower curtain, he is unable
to see objects clearly through the curtain. But, if he looks over the curtain into a
distant mirror, he will see quite clearly his hand near the curtain. This difference
can by shown theoretically by varying the object positions in the two-lens, thin-lens
equation.
4.4.1 Verification with Adaptive Optics

Chapter 3 showed that adaptive optics could improve effectiveness of ground-
based laser weapons aimed at ICBMs and satellites. Since seeing is a small prob-
lem for downward-looking satellites, it is not necessary to enhance reconnaissance
cameras with adaptive optics. However, adaptive optics is important for observing
reentry vehicles or satellites with ground-based telescopes. Adaptive optics is used
in this way at the Air Force Maui Optical Station (AMOS), which is at 3000-m alti-
tude, improving seeing and reducing atmospheric and cloud absorption. AMOS’s
location at 20◦ N latitude allows it to observe most satellites, with a resolution of
5 cm for objects at an altitude of 125 km. AMOS telescopes track RVs with velocities
of 5◦ /s, and accelerations of 4◦ /s2 . Adaptive optics is used with a 1.6-m telescope
to reduce visible seeing to the diffraction limit of 0.4 μrad. However, much better
photos can be obtained with cameras in space on Hubble or in a shuttle as they
approach objects.
P1: GFZ
4.5 Digital Image Processing

The development of electro-optic CCDs and very large scale integrated circuits has
enhanced ability to monitor military activities and arms control treaties. Digital
image processing restores and enhances photographs that are blurred by many
causes. Some approaches use the subtraction of one picture from another, enhance-
ment of edges and lines, removal of high-spatial-frequency noise, enhancement of
contrasts by the removal of clouds, and the search for patterns such as missile silos
or mobile missile launchers. These techniques are used on images obtained by pho-
tography, radar, sonar, infrared, and x-rays. The formation of a two-dimensional
image g(x,y) involves the integral of the image spreading function,

g(x, y) = h c (x − x1 , y − y1 )s(x1 , y1 )d x 1 dy1 + n(x, y). (4.7)
The composite point-spread function h c results from image motion, detector

distortions, and optical system aberrations. The direct and indirect flux at the sensor
is s and the noise is n. The information to be extracted from the image is difficult to
obtain because the point-spread function is multiplicative and the noise is additive.
For this reason, the analysis of the image is often carried out in the frequency domain
with Fourier transforms.
4.5.1 The Convolution Theorem

If we ignore the noise term, the Fourier transform to the frequency domain is
G(u, v) = H(u, v)S(u, v) (4.8)
where G, H, and S are the two-dimensional Fourier transforms of the image g(x,y),
the point-spread function h c , and the flux signal s(x,y), respectively. The beauty
of the convolution theorem is that the frequency domain equation is less com-
plex, since it is multiplicative without an integral. Consider a one-dimensional
photograph of light spots from two delta function lasers that are broadened by a
point-spread function,
g(x) = N[Aexp(−x 2 /2σ 2 ) + B exp(−(x − D)2 /2σ 2 )]. (4.9)
If we assume a Gaussian point-spread function,
h c = N exp(−(x − x1 )2 /2σ 2 ) with N = (2π σ 2 )−1/2 , (4.10)
and two delta functions for the lasers, the image function is

g(x) = N [exp(−(x − x1 )2 /2σ 2 ][Aδ(0) + Bδ(D)]d x 1 . (4.11)
Since this integral recovers the measured image g(x), the choice for h c and the laser
intensities/locations are correct. The point-spread function h c can be measured in
orbit by shining a laser from the ground to the satellite. The laser-spot object g and
the measured signal s are Fourier transformed to G and S. These are combined to
P1: GFZ
4.5. Digital Image Processing 85
H = G/S, which is Fourier-transformed to real space to give h c . This function in-

cludes not only the point-spread function of the optical system but also atmospheric
seeing effects, which vary with time.
4.5.2 Fourier Addition Theorem

A photo of a silo can be hidden by lack of contrast and photographic noise. When
optical resolution is small compared to silo size, the point-spread function can be
ignored, giving
g(x) = s(x) + n(x). (4.12)
If the silo diameter is much larger than the average noise wavelength, much of
the noise can be removed in the frequency domain. The Fourier transform addition
theorem is
F [g(x)] = G(k) = F [s(x) + n(x)] = S(k) + N(k), (4.13)
where F is the Fourier transform operator. As a pedagogical example, we add

random noise to a box function silo to obtain
g(x) = NN Rnd(1) + s(x), (4.14)
where Rnd(1) is a random number between 0 and 1, NN is the maximum value of

the noise, and s(x) is a box function representing the silo in Fig. 4.1.
Figure 4.1. Fourier-transformed image. (a)

The image of the silo g(x) is severely hid-
den by a large signal-to-noise ratio S/N =
1 (NN = 2 times the silo height). The silo
is 20 pixels wide in a field of 100 pixels.
(b) The spatial data is Fourier-transformed
to obtain its frequency spectrum: F [g(x)] =
G(k). (c) Most of the noise is removed from
the photo by Fourier-transforming only the
four lowest frequency components back
to position space to make the silo more
apparent.
P1: GFZ
4.6 Infrared Reconnaissance

Infrared detectors are made with gallium impurities in silicon which have energy
levels that can be excited by IR at 0.1 eV above the valence band. IR cameras obtain
thermal images of objects from orbit, which are very useful for finding covert
facilities and weapons, as well as missile and weapon tests. IR reconnaissance can
detect temperature differences of 0.1 K, but their spatial resolution is not as sharp
as visible cameras. The longer wavelengths of IR increase diffraction-broadening,
and the continuous distribution of heat radiation provides less information than
visible spectra, but IR has sufficient resolution to detect silos and vehicles. Lastly,
IR arrays have tens of thousands pixels as compared to optical CCDs with arrays of
millions of pixels. Two atmospheric wavelength windows transmit IR, one between
3–5 μ and the other between 8–14 μ. The blackbody radiation maximum for room
temperature (300 K) objects is at
λmax = 2897/T = 10 μ, (4.15)
which is conveniently located in the center of the 8–14 μ window. Space-based
photography prefers the 8–14 μ window because Earth is at 300 K and the window
is far from the sun’s reflected near-infrared rays. See Fig. 4.2. The 10-μ region has
Figure 4.2. Electromagnetic radiation at different temperatures. The radiation from the sun
at 6000 K extends from the ultraviolet (UV) to the infrared (IR), peaking in the visible (VIS)
region. Missile plumes with CO2 and H2 O combustion products are readily detectable at
geosynchronous orbit in the short- and midwave infrared (SWIR and MWIR). Colder bodies,
such as reentry vehicles, decoys, booster bodies, satellites, and the Earth radiate in the
longwave infrared (LWIR). Different detectors are required to detect the various objects
(SDI Technology, Office of Technology Assessment, 1988).
P1: GFZ
4.6. Infrared Reconnaissance 87
poor resolution since diffraction-broadening is 20 times greater than that of light

(10μ/0.5μ). IR angular resolution from a 2-m mirror is
θ = 1.2λ/D = 1.2(10 μ/2 m) = 6 μrad, (4.16)
giving a spatial resolution of
h o = oθ = (150 km)(6 μrad) = 0.9 m. (4.17)
This resolution is similar to that of the 4-m Kitt Peak telescope, which observed a
planet 5 μrad from the star VB-8. Adaptive optics is not used on IR systems since
atmospheric broadening is usually less than diffraction broadening.
4.6.1 IR Temperature Sensitivity

Room temperature IR peaks at 10 μ, which is at the center of the 8–14 μ win-
dow. This causes the window response of IR detectors to be proportional to the
fourth power of temperature, T 4 . The difference in accumulated charge (voltage)
for objects separated by a small temperature difference (T) is proportional to the
radiation difference at the two temperatures, or
(volts) = (σ T 4 ) = 4σ T 3 T = (σ T 4 )(4)(T/T) (4.18)
when T T. The voltage accuracy needed to measure with an accuracy of T =
0.1 K is
volts/volts = 4(T/T) = 4(0.1 K/300 K) = 10−3 . (4.19)
Absolute temperature can be obtained from voltages in two or more IR windows,
such as the ratio of the 3–4 μ window to the 8–14 μ windows.
4.6.2 IR Detection of RVs and Satellites

Defense systems use the 8–14 μ window when searching for cool RVs. Early warn-
ing satellites use the 3–5 μ window to look for hot missile plumes at 2.7 and 4.3
μ from CO2 and H2 O. Can one observe a 1-m2 satellite or RV at 1000 km? A nu-
merical integration of the 273-K Planck radiation distribution shows that 35% of
the radiation lies within the 8–14 μ window. A 1-m2 , ice cold, blackbody surface
radiates with a power
Pir = 0.35(σ T 4 A) = 0.35(5.7 × 10−8 )(2734 )(1) = 110 W (4.20)
in the 8–14 μ window. At a distance of 1000 km (R), a 1-m diameter (D) mirror
collects
P = Pir π (D/2)2 /4π R2 = 7 × 10−12 W. (4.21)
Freezing temperature of 273 K gives IR peaked at 11 μ, or 0.11 eV, which gives a
photo emission rate of
d Nγ /dt = P/0.11 eV = (7 × 10−12 W)(1 eV/1.6 × 10−19 J)/0.11 eV = 4 × 108 /s.
(4.22)
P1: GFZ
The IR photons are focused with a mirror onto Si:Ge which has a quantum
efficiency of 0.5 in the IR to give 2 × 108 electrons/s. An Si:Ge detector has a
detectivity of
D = (S/N)/P = 5 × 1012 /W, (4.23)
where S/N is the signal-to-noise ratio and P is the IR power in the 8–14 μ window.
This value of D was multiplied by a factor of 100 to account for detector size,
amplifier bandwidth, and field of view. This gives a signal-to-noise ratio of
S/N = DP = (5 × 1012 )(7 × 10−12 ) = 35, (4.24)
which allows good detection in short periods of time.
4.6.3 Monitoring Orbital Nuclear Power

Space-based weapons need considerable continuous and burst electrical power. To
constrain competition of space-based weapons, a group of US and Soviet scientists
proposed a ban on nuclear power in Earth orbit. Monitoring for the presence of
nuclear power in orbit would be an important aspect of this treaty.
We determine the ability of Air Force Maui Optical Station (AMOS) to observe
the American 100-kWe space reactor (SP-100). The SP-100 was designed to develop
power with η = 4% efficiency, releasing rejected infrared power of
PIR = Pelec [(1/η) − 1] = 0.1 MW[(1/0.04) − 1] = 2.4 MWt . (4.25)
The 90-m2 thermal radiators rise to 827 K to radiate the 2.4 MWt . The SP-100
irradiance at 1000 km is
i IR = f P IR /4π R2 = 0.15(2.4 × 106 W)/4π (108 cm)2 = 3 × 10−12 W/cm2 , (4.26)
where IR fraction f in the 8–13 μ window at 827 K is 15%. This can be readily
detected by AMOS which can detect i limit = 5 × 10−18 W/cm2 in the 8–13 μ window
with a signal-to-noise of 1 after the background subtraction. (i limit = 10−17 for the
3–4 μ window). Since the detection limit is 10−6 smaller than the SP-100 signal at
1000 km, AMOS can detect SP-100 to 1,000,000 km, four times the distance to the
moon.
4.7 Radar Reconnaissance

4.7.1 Synthetic Aperture Radar (SAR)
SAR systems on satellites obtain images to monitor arms control treaties, while
Earth-based radar monitors ballistic missile tests and obtains photos of missiles
and RVs. Because of radar’s long wavelength, SAR resolution cannot compete
with optical systems. However, radar is used at night and it locates objects through
clouds and rain, as well as in wood buildings. Radar’s long wavelength requires
an antenna several miles in length to obtain good resolution. Such an antenna is
P1: GFZ
4.7. Radar Reconnaissance 89
created by using the satellite’s motion to increase the effective size of the antenna. A
resolution of 25 m has been obtained with the SEASAT radar satellite at an 800-km
altitude, but resolution of 1–3 m is obtained at 10 GHz and lower altitudes. SAR’s
angular resolution θ is determined through a calculation involving the effective
antenna diameter, which is twice the product of the satellite velocity and the time
SAR data are received for computer processing:
θ = λ/2vt process = (0.03 m)/2(7500 m/s)(0.2 s) = 10 μrad, (4.27)
where λ = 0.03 m at 10 GHz, orbital velocity v is 7.5 km/s, and processing time
tprocess is 0.2 s. Then, for a slant range R of 200 km, SAR has a spatial resolution of
R = Rθ = (2 × 105 m)(10 μrad) = 2 m. (4.28)
By combining 2 SAR images, subsidence of 2 mm can be detected using interfer-
ometric synthetic aperature radar (InSAR).
4.7.2 Large Phased-Array Radar

Large phased-array radar (LPAR) is located at the Cobra Dane radar station in the
Aleutians and the Cobra Judy on the Observation Island ship. Cobra Dane operates
at 200 MHz, giving a time-range resolution of λ = c/ f = 1 m. However, reflected
radar at 30◦ off the normal to the radar plane (the boresite) gives a disparity in
resolution on the radar plane,
Rplane = (30-m diameter)(sin 30◦ ) = 15 m (4.29)
between extreme points on a wavefront. To remove this spread, Cobra Dane has 96
subarrays to measure different cycles of the wave, and these signals are combined
electronically. The 96-array pattern has a diameter of 10 arrays, reducing R to
about 1 m, consistent with the time-range resolution.
Reentry vehicle size is measured with ground-based radar using inverse syn-
thetic aperture radar (ISAR), in which the moving and rotating RV supplies an-
tenna motion. The Doppler shift difference of 10-GHz radar between the two ends
of a tumbling RV is
f = 2 f ωL/c = 2(1010 Hz)(1 rad/s)(2 m)/(3 × 108 m/s) = 140 Hz, (4.30)
which is doubled for radar reflection from a moving body. The RV tumbles at
ω = 1 rad/s, RV length L is 2 m and c is speed of light. The Doppler shift spectrum
is divided into cells giving a size resolution of 10 cm.
4.7.3 Ballistic Missile Coefficient

START limits changes in throw-weight on new types of missiles because extra weight
increases the capacity to add reentry vehicles. RV mass and accuracy are measured
with Doppler shift measurements that observe RV motion, including the effects of
atmospheric drag. Small drag implies good aerodynamic shapes for good accuracy
and heavy mass.
P1: GFZ
As a pedagogical aside, the measurement of descent time of a coffee filter demon-

strates the validity of a velocity-squared drag force. The low-velocity differential
equation for an object falling in constant density air is
md 2 y/dt 2 = mg − K v2 , (4.31)
where K is a constant. Since coffee filters are light, they quickly arrive at terminal
velocity vt = (mg/K )1/2 , with a time to fall a height H approximately equal to
tdrop = H/vt . This is verified by timing the descent of stacked filters with varying
numbers of filters from the same height. For example, a one-filter drop versus a
four-filter drop (m4 = 4m) gives a ratio of drop times,
tdrop-1 /tdrop-4 = v4 /v1 = (m4 /m1 )1/2 = 2, (4.32)
indicating the heavier filter falls twice as fast. Of course, measurement of reentry
vehicle mass is much more complicated as it travels with three-dimensional motion
in a varying gravitational field with variable air density. Nonetheless, the constant
K can be simplified to
K = ρC d A/2, (4.33)
where ρ is air density, Cd is the drag coefficient, and A is RV lateral area. Time-
dependent Doppler shift data provide the RV deceleration profile, which deter-
mines the reentry vehicle drag coefficient and mass. The ballistic missile coefficient
is the ratio of RV weight to drag area, which is the ratio of good ballistics divided
by unwanted drag. An RV’s ballistic missile coefficient for a system with m = 500
kg, A = 1 m2 , and Cd = 0.15 is
β = mg/Cd A = (500 × 10 N)/(0.15)(1 m2 ) = 30,000 Pascals. (4.34)
4.8 Nuclear Tests in Space

The CTBT (Section 4.2) calls for a ban on all nuclear tests, anywhere and for all
time, but it was defeated in the US senate in 1999. The CTBT would slow the devel-
opment of new types of nuclear weapons, since it bans nuclear weapons testing.
The CTBT could improve the political climate among nuclear weapon states, and
its adoption is consistent with the Nuclear Non-Proliferation Treaty (NPT). The
CTBT debate centered on its verifiability, the need it posed to develop new nuclear
warhead designs and the ability of nuclear weapon states to maintain safe and
reliable nuclear weapons. It also considered what violating nations might gain by
cheating on the treaty. In this section we examine detection of nuclear weapons
tests in space.
The LTBT of 1963 forbids testing of nuclear weapons in space, in the atmosphere
and underwater. LTBT monitoring must be able to detect nuclear debris from an ex-
plosion above the Earth. We estimate the neutron, x-ray, and prompt γ -ray fluences
from weapons exploded in space at a distance of 20,000 km, where GPS satellites
reside. The results given below agree with results from the Los Alamos Vela satellite
P1: GFZ
4.8. Nuclear Tests in Space 91
program of the 1960s, which showed these signals to be readily observable with
satellite detectors.
4.8.1 Neutrons
The yield of a 1-kton weapon in MeV is (4.2 × 1012 J/kton)(1 MeV/1.6 × 10−13 J) =
2.6 × 1025 MeV. If one neutron per fission escapes the warhead, a 1-kton weapon
releases Nn neutrons according to
Nn = (2.6 × 1025 MeV)(1 n/170 MeV) = 1.5 × 1023 n. (4.35)
The neutron fluence fln from a yield of Y kton at a distance of R km is
fln = Y × 1.5 × 1023 /4π (R × 105 cm)2 ≈ 1012 Y/R2 n/cm2 . (4.36)
The particle fluence from a 1-kton explosion at a distance of 20,000 km from a
GPS is
fln-GPS = (1012 )/(4 × 108 ) = 2500 n/cm2 -kton. (4.37)
4.8.2 X-rays
About 70% of yield in space appears in the form of x-rays. This gives an x-ray
energy fluence flenergy-x at a distance of R km from a yield of Y kton of
flenergy-x = (0.7)(Y × 2.6 × 1025 MeV)/4π (R × 105 cm)2

= 1.4 × 1014 Y/R2 (MeV/cm2 ). (4.38)
If the radiating fireball has a temperature of 20 million K, the average x-ray energy
is (with kB T = 1/40 eV at 300 K)
E av = 3kB T = (3/40 eV)(20 × 106 K/300 K) = 5 keV. (4.39)
The x-ray particle fluence at a GPS detector at 20,000 km is
flx-GPS = flenergy-x /E av = (1.4 × 1014 Y)/(0.005 MeV)(2 × 104 km)2 = 108 Y/cm2 .
(4.40)
4.8.3 Prompt Gamma Rays

About 0.3% of yield appears as prompt gamma rays, giving an energy fluence of
flenergy-γ = (0.003/0.7)(1.4 × 1014 )Y/R2 = 6 × 1011 Y/R2 (MeV/cm2 ). (4.41)
The average energy of prompt γ -rays is about 1 MeV, giving a γ -ray fluence at
GPS of
flγ -GPS = (6 × 1011 MeV/cm2 -kton)Y/(1 MeV)(2 × 104 km)2 = 1500 Y-γ/cm2 .
(4.42)
P1: GFZ
4.9 Atmospheric Nuclear Tests

Atmospheric nuclear explosions give unique double-peaked optical signatures.
The initial burst of x-rays lasts less than 1 μs, after which they are absorbed in the
first few meters of air, creating a million-Kelvin fireball of heated air. Shortly after-
wards a shock wave carries the energy beyond the fireball. Since high-temperature
shocked gas is ionized, it is opaque to the fireball optical radiation, thus reduc-
ing luminosity of the fireball. After the shocked gas has expanded and cooled,
it again becomes transparent to light and the fireball optical intensity again rises
for a second pulse. Optical “bhangmeters” detect double pulses above the intense
brightness background of the earthshine. Empirical algorithms give the yield of
weapons from delay times of the first minimum intensity and from the second
maximum intensity. See Fig. 4.3.
The time from explosion to the minimum light intensity is determined as the
shocked region expands beyond the fireball, blocking fireball radiation, until fur-
ther expansion cools the shock and unblocks the radiation. Shock front velocity is
not too different from velocity of sound at high pressures and temperatures. The
shocked region from a 20-kton explosion has a temperature of some 100,000 K dur-
ing the millisecond regime, giving a shock velocity (pegged to sound velocity) at
breakaway when the intensity begins to rise,
vshock = vsound = (331 m/s)(100,000 K/300 K)1/2 = 6000 m/s. (4.43)
The radius of the fireball at breakaway is R = 520Y with R in meters and Y in
0.4
megatons. The time to the intensity minimum at breakaway is

tbreakaway = 520Y0.4 m/6000 m/s = 87Y0.4 ms, (4.44)
Figure 4.3. Optical bhangmeter. The characteristic double-peak signal is from a 19-kton
atmospheric nuclear test. The optical photometers record the luminosity of the fireball as a
function of time (Argo, 1986).
P1: GFZ
4.10. Underground Nuclear Tests 93
Figure 4.4. EMP detection. A

simulated EMP signal is seen
by a satellite receiver tuned to
50 MHz with a 2-MHz band-
width. The ionosphere disperses
the signal as a function of the fre-
quency complicating the inter-
pretation of the received signal.
The second peak is a ground re-
flection (Argo, 1986).
in fair agreement with Brode’s empirical relation, tbreakaway = 60Y0.4 ms. For a 19-
kton weapon, t = (110 mec)/(6000 m/sec) = 18 msec, in good agreement with
Fig. 4.3.
Additional confirmation of an explosion is obtained from the coincidence of a
bhangmeter signal and arrival of an electromagnetic pulse (EMP problems 1.17
and 1.18) which accompanies nuclear explosions. See Fig. 4.4. Additional evidence
is available from the GPS system that observes any spot on Earth with 4–8 of its
24 global positioning satellites satellites. Accurate GPS clocks allow triangulation
for accurately determining the position of an atmospheric nuclear explosion. Had
these instruments functioned in 1979, they might have removed ambiguity from
the possible nuclear explosion “event” over the South Atlantic. An independent
panel of the Presidential Office of Science and Technology Policy reported in 1980
that the signals were probably not from a nuclear explosion, but other scientists
still disagree.
4.10 Underground Nuclear Tests

Nuclear tests have been confined to underground locations by the United States
and Russia since 1963, as well as by UK (1958), France (1974), China (1980), India
(1974, 1998), and Pakistan (1998). Seismographs are the primary tool to monitor
underground tests, but other technologies supplement seismic data. The seismic
traces from nuclear explosions differ from earthquake traces in several ways. Nu-
clear explosion seismographic data display higher frequency components because
the duration of explosions is shorter as compared to earthquakes. The ratio of the
P1: GFZ
Figure 4.5. IMS seismic monitoring limit (tons). Projected 90% probable, 3-station detection
thresholds in mb seismic magnitude units for the IMS network of 50 primary stations (DoD,
2002).
short-period, pressure body wave magnitude mb to the long-period, surface wave mag-
nitude MS is larger for weapons than for earthquakes. The CTBT IMS will have the
capability to monitor explosions with high confidence (90% certainty) to a seismic
mb level of 3.5, which corresponds to a tamped explosion of about 0.1 kton in hard
rock throughout Eurasia and North Africa. See Fig. 4.5 and 4.6. This is better than
the 1 kton that was originally projected for the IMS, an assessment that was too
cautious in that it did not take into account the growing number of close-in, re-
gional stations. A neighboring state could place regional seismographs close to a
suspected region to improve monitoring. Finally, chemical explosions are usually
identifiable because they are not spherical explosions, but rather ripple-fired along
a line to reduce costs. Voluntary notifications for chemical explosions larger than
0.3 kton can reduce suspicions about chemical explosions.
The IMS is deploying seismographs without the CTBT formally entering into
force. The IMS will consist of 50 primary and 120 auxiliary seismic stations, as well
as 60 infrasound stations (1-kton global atmospheric threshold detection), 11 hy-
droacoustic stations (less than 1-kton global oceanic detection), and 80 radionuclide
stations (less than 1-kton, global atmospheric detection). In addition, the United
States will use satellite optical bhangmeters, particle detectors, and EMP detectors
to monitor atmospheric tests. Lastly, US NTM of satellite reconnaissance, human
intelligence (humint), and other “ints” will combine to make intelligence gathering
greater than the sum of its parts. A nation’s fear of being spotted by the IMS and
NTM deters it from cheating, and these measures will be buttressed by OSIs. See
Fig. 4.7 and 4.8.
P1: GFZ
Figure 4.6. IMS seismic monitoring limit (tons). Seismic magnitudes of Fig. 4.5 are converted
to yields in tons for the IMS network of 50 primary stations. The IMS detection threshold is
about 0.1 kton for most of the northern hemisphere, below 0.5 kton for most of the world,
and below 0.01 kton at Novaya Zemlya. The IMS system with 33 stations detected 0.025-kton
explosions at the Semipalatinsk Test Site (DoD, 2002).
Figure 4.7. IMS infrasound monitoring limit (kiloton). Projected 90% probable two-station
detection thresholds for atmospheric explosions for the planned IMS network of 60 infra-
sound stations. Thresholds are below 0.5 kton on continents in the northern hemisphere and
below 1 kton world wide. A space shuttle launch from Cape Canaveral was readily detected
by a prototype infrasound station near Los Alamos, New Mexico at a distance of 1600 miles
(DoD, 2002).
P1: GFZ
Figure 4.8. IMS radionuclide monitoring limit (kiloton). Probability of one-station detection
within 5 days of a 1-kton atmospheric explosion by the planned 80-station IMS radionuclide
network. Probabilities of detection exceed 90% across most of Eurasia and exceed 50% over
most of the southern oceans. A prototype system detected radionuclides released from a
steel recycle plant with a source level of a 0.001-kton explosion at a distance of 1200 miles
(DoD, 2002).
4.10.1 Explosion in a Cavity

There are little data on nuclear tests in cavities. A fully decouped test needs a cavity
sufficiently large to minimize the observed yield by, at most, a factor of 70. The only
fully decoupled test took place in 1966 when the 0.38-kton Sterling explosion was
exploded in a Mississippi salt cavity with a 17-m radius. The cavity was created by
the 5.3-kton, Salmon explosion. The Soviets carried out a 9-kton test in a cavity at
Azgir in 1976. This test was only partially decoupled, as the weapon was too large
for the cavity’s 36-m radius, itself created by a 64-kton previous test. If a nuclear
weapon is placed in a cavity of sufficient size, the blast pressure on the cavity wall
will fall below the material’s elastic limit, reducing the seismic signal strength by
a theoretical factor of 7 at 20 Hz and 70 at lower frequencies.
Covert testing is complicated by possible radioactive venting that can be de-
tected. The Soviets had 30 percent of its tests vent, and the United States had severe
venting problems during its first decade of underground testing. Smaller tests are
harder to contain than larger ones, as the last four US tests that vented had yields
of less than 20 ktons. It is hypothesized that smaller explosions do not sufficiently
glassify cavities, and they do not rebound sufficiently to seal fractures with a stress
cage, thus making them easier to vent radioactivity. In addition, a nation new to
testing is not likely to have adequate knowledge to predict the test yield, and thus
rule out yield excursions that exceeded the decoupling range.
P1: GFZ
If blast pressure exceeds the elastic limit of the cavity’s wall material, sufficient
energy is absorbed to crack the wall, increasing coupling to the wall and increasing
the seismic signal. Critical cavity size depends on explosion depth, but it is usually
assumed to be about 0.5–1 km. The critical radius for decoupling is
Rc = (20 m)Y1/3 , (4.45)
with Y in kilotons. From this, a 30-kton explosion needs a cavity radius of 60 m
(a 20-story building) to achieve full decoupling—an extraordinary engineering
challenge when one considers the secrecy requirements. One expects that Rc is
proportional to Y1/3 since the energy to fill the volume of the cavity to a critical
pressure is proportional to the yield, or Rc3 αY. We will estimate the 20 m coefficient
from first principles for a 1-kton blast in salt. It is easier to clandestinely mine a salt
cavity using water solvents than to mine granite cavities, but only a few nations
have salt deposits thick enough to pull off this kind of violation.
Because an explosion occurs very rapidly, an adiabatic expansion results with
PVγ = C, a constant. The yield Y to compress air to the elastic limit of salt is

Y = − PdV = − CV −γ dV = CV 1−γ /(γ − 1)
= Po (4π Rc3 /3)/(γ − 1) = Po Vc /(γ − 1), (4.46)

where Po is the wall elastic limit and Vc is minimum cavity volume. Using Y =
1 kton, γ = 1.2 (very hot air), and Po = 440 bar for salt’s elastic limit, we obtain a
minimum elastic radius of Rc = 16 m, which is consistent with the 20 m value.
4.10.2 Threshold Test Ban Monitoring

Monitoring the TTBT requires that yield be measured to determine if an explosion
exceeded the treaty’s 150-kton threshold. Because the plate below the Nevada site
is young, its seismic waves are diminished more than waves from the Shagan
River Test Site in Kazakhstan, which resides on a much older plate. Because of
this difference, US explosions appear smaller compared to explosions of the same
yield at the Soviet site. Ignoring this bias difference, the United States incorrectly
charged the Soviets with a “likely” violation of the TTB treaty. The magnitude of
an explosion is calculated from the body wave seismic magnitude (the pressure
wave) at
mb = a + b + c log Y, (4.47)
where mb is the magnitude for a 1-Hz body wave, a is the 4.1 magnitude of a 1-kton
explosion, b is the bias correction for a test site, c is the slope of 0.74, and Y is yield
in kiloton. A 150-kton yield at the Nevada Test Site has a mb value of
mb = 4.1 + 0.74 log 150 = 4.1 + 1.61 = 5.71, (4.48)
while a similar 150-kton explosion at the Soviet site with a bias of 0.4 is
mb = 4.1 + 0.4 + 1.61 = 6.11. (4.49)
P1: GFZ
The United States initially assumed there was no bias between the two sites (b =
0), which gave the United States a false impression that the Soviet’s explosion at
6.11 mb was a serious violation of the yield limit at
Y = 10[(6.11−4.1−0)/0.74] = 520 kton. (4.50)
Over time, US estimates of the bias difference rose as government scientists

learned geophysics. If a bias b = 0.2 is used with mb = 6.11, Y would be 280 kton.
If b = 0.3 then Y = 200 kton, and if b = 0.4 then Y = 150 kton. These differences
show that choice of bias is exceedingly important in judging compliance. Security
classification of seismic data prevented a thorough discussion by experts. Sectors
of the US policy community purposely ignored geological bias to maintain the
“likely violation” charge against the Soviets.1 The designation of “likely violation”
on nuclear testing greatly hindered negotiations on the CTBT in particular and
arms control in general.
4.10.3 CTBT Monitoring Limits

The CTBT treaty is monitorable to about 0.1 kton for tamped explosions in hard
rock in all of Eurasia. The National Academy of Sciences convened a bipartisan
panel of experts to carry out a study during 2000–2002, Technical Issues Related to
the Comprehensive Nuclear Test Ban Treaty. The panel concluded, “the only evasion
scenarios that need to be taken seriously at this time are cavity decoupling and
mine masking.” Successful covert testing involves at least seven issues, each with
a differing probability of success:
1. Violators need excellent yield estimates to avoid yield excursions.

2. Violators need to hide removed materials from satellites.
3. Crater and surface changes from testing are observables.
4. Radioactive releases from tests often observed in former USSR.
5. Regional signals at 10 Hz improve detection.
6. A series of tests is needed to develop significant weapons.
7. Human and other intelligence can give information.
Because the net success probability for hiding a covert test in a cavity is the
product of the individual success probabilities, the NAS panel did not use a
decoupling factor of 70 times the 0.1-kton limit to obtain a maximum cheating
limit of 7 kton. Rather, it concluded the following: “Taking all these factors into
account and assuming a fully functional IMS, we judge that an underground
nuclear explosion cannot be confidently hidden if the yield is larger than 1 or
2 kton.”
1
The author served as technical lead on TTBT issues in the State Department (1987) and the
Senate Foreign Relations Committee staff (1990–92).
P1: GFZ
4.11. How Much Verification Is Enough? 99
4.10.4 CORRTEX Monitoring

An alternative method, the Continuous Reflectometry for Radius versus Time Ex-
periment (CORRTEX), was developed at Los Alamos to monitor the TTB treaty.
It was used only once in this capacity in 1988 to refine analysis of the bias factor
in 1988. It is inferior to seismology in many respects and expensive as well. In
CORRTEX, electrical pulses are sent down a cable and reflected back to the sur-
face from the end of the cable. When the cable is located near a nuclear explosion,
the crushing of the cable shortens the cable and the reflection time. The reflection
time of the pulses gives a measure of the shock wave radius rt as a function of
time,
rt = [l 2 + (do − dt )2 ]1/2 , (4.51)
where l is the distance between the weapon emplacement hole and the CORRTEX-
offset emplacement hole, do is the depth of the explosion, and dt is the time-
dependent depth of the shock front as it intersects the CORRTEX cable. CORRTEX
and seismology are both indirect techniques that must be calibrated with al-
gorithms (problem 4.17) that describe coupling between yield and geological
media.
4.11 How Much Verification Is Enough?

Nations need to quantify national security threats in order to determine if an arms
control treaty can indeed be “effectively” verified. The “effective” standard would
require that the United States have sufficient, timely warning to respond to an attack
from covert warheads that could significantly damage US national security beyond
what could be done by the already legal former-Soviet strategic forces. Verification
would be carried out through a variety of monitoring inspections. An important
one for START is the re-entry vehicle onsite inspection (RVOSI), which determines
if more than the allowed number of RVs is present on ICBMs, SLBMs, and bombers
(Section 4.2). The decision is often made on the basis or observing whether a bump
is filled in a cloth covering the RV bus. Since it would be difficult to have more
than one nuclear warhead in an RV, the observation of a filled-bump is deemed to
be due to a warhead. START II would have raised the number of RVOSIs from 10
to 14 per year because of the possible danger, for example, of adding one or two
small warheads to the single-warhead SS-27 or by up loading formerly MIRVed
Minuteman IIIs and SS-19s.
4.11.1 START II Breakout

In spite of the fact that START II is moribund, it is useful to examine a treaty that
was declared verifiable by the US intelligence community, the US Senate, and the
Russian Duma. The following calculations assume a robust Russia, with the ability
to covertly upload 1500 warheads by returning the SS-19 from a single warhead
P1: GFZ
to its capacity of six warheads and to increase the SS-25/27 from one warhead to
three warheads:
ICBMs Covert Warheads/ICBM Covert Warheads

105 SS-19 6−1=5 525
500 SS-25/27 3−1=2 1000
Random selection assumes the inspecting party has no knowledge as to where a

violation is most likely to occur. In practice, the intelligence community gets myriad
data to guide the selection of RVOSI sites, an advantage that improves detection
beyond what is possible with random choices. If Russia maintained all 105 SS-19s
and uploaded extra RVs on 10 of them, the probability of detection from one RVOSI
per year would be
Pdetect-1 = V/S = 10/105 = 0.1 = 10%, (4.52)
where the number of violations V is 10 and the number of sites S is 105. If n RVOSIs
were carried out each year, the annual detection probability would be one minus
the probability of nondetection, or
Pdetect-n = 1 − Pnondetect-n = 1 − [1 − Pdetect-1 ]n = 1 − [1 − V/S]n . (4.53)
For example, if three RVOSI’s were performed, the annual probability of detection
would not be 3 × 10% = 30%, but rather it would be
Pdetect-3 = 1 − [1 − 10/105]3 = 0.26 = 26%. (4.54)
Thus, additional inspections are marginally less effective, per inspection. We allot
the 14 RVOSIs as follows: 3 for SS-19, 6 for SS-25/27, 4 for SSBN, and 1 for heavy
bombers.
4.11.2 High Confidence

The intelligence community defines “high confidence” verification as a 90% proba-
bility (within 2 standard deviations, or 2σ ) of detecting a violation. In our case this
would mean a violation in the declared Russian arsenal of ICBMs and SLBMs. For
simplicity, we demand high confidence of detecting a violation of either the SS-19
or the SS-25/27, but with the same probability. The 9 annual RVOSIs dedicated to
both systems would give a detection probability on the two systems of
Pdetect (19 and 25/27) = 1 − [1 − Pdetect (19 or 25/27)]2 = 0.90. (4.55)
This would require 68% confidence (that is medium confidence, or 1σ ) of detecting

a violation for either system, that is
Pdetect (19 or 25/27) = Pdetect (19) = Pdetect (25/27) = 0.68, (4.56)

P1: GFZ
Problems 101
since 0.90 = 1 − (1 − 0.68)2 . For the SS-19s, the number of violations needed for a
68% confidence level for three RVOSIs on 105 sites is determined from
Pdetect (3 RVOSI on SS-19s) = 1 − (1 − V/105)3 = 0.68, (4.57)
which gives V = 33 violations, for 33 × 5 = 165 covert warheads. For the SS-25/27,
the number of violations needed for a 68% confidence level for 6 RVOSIs on 500
sites is obtained from
Pdetect (6 RVOSI on SS-25/27) = 1 − (1 − V/500)6 = 0.68, (4.58)
which gives 86 violations and 86 × 2 = 172 covert warheads. The total number of
covert warheads under the criterion of high confidence detection is, therefore,
covert warheads = (33)(5) + (86)(2) = 165 + 172 = 337. (4.59)
A violation of 337 warheads is a 10% violation on the treaty limit of 3500 war-
heads. The additional strategic damage caused by such a violation, beyond the dam-
age to US strategic forces from the legal base case of 3500 warheads, is marginal since
the first 3500 warheads had already reached a point of “diminishing return.” The
diminishing probability/warhead of destroying a silo for additional warheads is analogous
to the diminishing probability/inspection of detecting a violation with further inspections.
(Section 2.6 and problems 4.19–21.)
Problems
4.1 Unratified treaties. A country has signed but not ratified a treaty. Should this
country be held responsible for compliance to the terms of the treaty? What
are examples of arms control treaties that were not ratified, but were complied
with.
4.2 Moving satellite film. How fast should film be moved to remove motion-blur
from a 7.5 km/s satellite at 150-km altitude with a 6-m focal length mirror.
4.3 CCD reconnaissance. Some day reconnaissance satellites might have 10-m
diameter mirrors and 0.5-μ CCD pixels. (a) What is resolution θ for geomet-
rical and diffraction-broadening at a 150-km altitude? (b) What is the spatial
resolution of this system in centimeters?
4.4 Experimental seeing. (a) You are inside a shower with your hand placed just
inside a translucent shower curtain. Look over the top of the curtain to see
observe your hand in a distant mirror. How does this image of your hand
compare to the image of your hand when it is placed just outside the curtain
while you remain in the shower? What can you conclude? (b) Obtain a real
image 5 m from a wire mesh with a 0.5-m lens. What happens to the image
when a bunsen burner is placed near the wire mesh as compared to near the
lens. (c) Explain the results for satellite cameras.
4.5 Theoretical seeing. Apply the thin lens equation (Section 4.3) to the case of
two separated, coaxial lenses. (a) For a land-based telescope let the object be
a long distance from a first lens that has a long focal length, representing a
P1: GFZ
refracting atmosphere. The second lens is the telescope with a shorter focal
length. Slightly shift the position of the first lens (the atmosphere) and find
the position of the final image. (b) For the reconnaissance case let the object
be near the long focal length atmosphere. Show that shifting the position of
the first lens (the atmosphere) has a much smaller effect on the position of the
final image for this case.
4.6 Composite spread function. What happens to the integral image-spreading
function g(x,y) if the composite point-spread function h c is a delta function?
What is the image function g(x) for the example of two lasers with this point
spread function?
4.7 IR from hot RV. An RV with a surface area of 1 m2 enters the atmosphere and
its temperature rises to 1000 K. (a) What is the radiative power of the RV? (b)
What is the predominant IR energy and wavelength? (c) What fraction of the
energy is radiated in the 10–14 μ window?
4.8 SAR resolution. What is the radar frequency of a synthetic aperture radar that
obtains a 1-m resolution at 150 km with a processing time of 0.2 s?
4.9 LPAR sizing. The Doppler shifts are detected from a tumbling RV at +80 Hz
and −40 Hz from a 10 GHz signal. What is the configuration of the RV, which
is 2 m long and rotating at 1 rad/s?
4.10 Ballistic missile coefficient. (a) What are the drop times and terminal ve-
locities for 1–5 coffee filters released from a 2-m height? (b) What are the
coefficients n and K in the drag force Kvn ?
4.11 Muffled cavity tests. What is the cavity-decoupling radius for tests of 0.1, 1.0,
10, 20, and 100 kton? What are some difficulties for this kind of covert nuclear-
weapon testing? (c) What happens if the cavity is made into an ellipsoidal
shape?
4.12 High-frequency components. (a) Fourier transform a one-dimensional
Gaussian of time duration into the frequency domain. Discuss results in terms
of earthquakes and nuclear tests. (b) Since high frequency components are ab-
sorbed strongly by the Earth, observation of these frequencies requires their
detection by close-in regional seismographs. Using a damping force that is
proportional to velocity, show that the amplitude of waves drops exponen-
tially as a function of frequency.
4.13 One ton at GEO. What are the neutron, x-ray, and gamma-ray fluences at a
satellite in GEO orbit (40,000 radius) from 1-ton and 100-ton explosions just
above the atmosphere?
4.14 Cavity at depth h. To contain a cavity explosion there must be sufficient mass
above the cavity to resist outward pressure. Show that 50% of pressure from
overbearing mass equated to cavity pressure gives ρgh/2 = Pc = (γ − 1)Y/Vc
where ρ is mass density, h is the depth of the cavity, and Vc is the volume of
the cavity. Show that this gives a minimum cavity radius of Rc = (20 m)Y1/3 at
h = 1 km.
4.15 One kiloton at Nevada and Shagan River test sites. (a) What is the body
wave magnitude mb of a 1-kton explosion at NTS? (b) What is mb for 1 kton
at SRTS with a bias of 0.4? (c) What is the physical cause for the bias factor?
P1: GFZ
Bibliography 103
4.16 Large yields at NTS and SRTS. (a) What are the body wave magnitudes mb
values for 150- and 200-kton explosions at NTS and SRTS? (b) What are the
yields of explosions of mb = 6.0 and 6.2 at NTS and SRTS with a bias of 0.4?
4.17 CORRTEX. (a) Derive Eq. 4.51, rt = [l 2 + (do − dt )2 ]1/2 . (b) Neither CORRTEX
nor seismology directly measure yield since both are calibrated with empir-
ically developed algorithms. The Los Alamos group measured shock wave
radius as a function of time and yield, r (t,Y) = aY1/3 (t/Y)b , where r is in me-
ters, Y is in kilotons, t in milliseconds, a is 6.29 and b is 0.475. For a 100-kton
test, how long does it take for the shock wave to travel 5, 10, and 20 m? What
is the velocity of the waves at 5, 10, and 20 m? (c) Drilling holes causes errors
in calculating the distance l between the CORRTEX and weapon holes. Deter-
mine the fractional yield error (Y/Y) from fractional displacement-distance
errors of l/l of 1m/5m and 1m/10m.
4.18 Verification standard: START versus CTBT. (a) What does the effective veri-
fication standard require of START and CTBT? (b) Discuss the congressional
acceptance of START I–II and the rejection of CTBT in terms of the effective
verification standard.
4.19 Diminishing returns with more inspections. (a) Show that the probability for
discovering a violation with integral numbers of random inspections (Eq. 4.53)
can be transformed for continuous variables to Pdetect = 1 − e −nf , for the case
of nf 1 where n is the number of inspections and f is fraction of sites having
a violation. (b) Why is it plausible that successive inspections have a declining
utility? Why is this true in the equation above?
4.20 Inspections for high confidence. (a) Show that the equation in problem 4.19
can be rewritten to determine the number of inspections n needed for a prob-
ability of detection Pdetect-n = P as n = ln(1 − P)/ln(1 − f ). (b) If a cheater
has 250 violations at 1000 locations, how many inspections are needed to ob-
tain high confidence (P = 0.9)? How many are needed to obtain medium
confidence (P = 0.68)?
4.21 Militarily significant violations. Assume the United States has 2000 SORT
warheads comprised of 50% SLBMs (2/3 at sea and 1/3 in port), 25%
1-warhead ICBMs in silos, and 25% bombers. Assume 1500 Russian warheads
have an 0.8 kill probability against the silos. (a) How many US warheads sur-
vive a worst-case Russian attack within treaty limits? (b) Determine US sur-
vivable warheads with Russian violations of 250–1000 warheads. Are these
violations militarily significant?
Bibliography
Argo, H. (1986). Arms Control Verification, K. Tsipis, D. Hafemeister and P. Janeway (Eds.),
Pergamon, New York.
Center for Monitoring Research, Department of Defense, Technical Issues Related to the Com-
prehensive Nuclear Test Ban Treaty, National Academy of Sciences, National Academy Press,
Washington, DC.
P1: GFZ
Dunn, L. (1980). Arms Control Verification and the Role of On-Site Inspection, Lexington Press,
Lexington, MA.
Elachi, C. (1987). Introduction to the Physics and Techniques of Remote Sensing, Wiley, New York.
Drell, S. and R. Purifoy (1994). Technical issues of a nuclear test ban, Ann. Rev. Nucl. Particle
Sci. 44, 285–327.
Fetter, S. (1988). Towards a Comprehensive Test Ban, Ballinger, Cambridge, MA.
Fetter, S., et al. (1990). Gamma-ray measurements of a Soviet cruise-missile warhead, Science
248, 828–834.
Graham, T. (2002). Disarmament Sketches: Three Decades of Arms Control International Law,
University of Washington Press, Seattle.
Jeanloz, R. (2000). Science-based stockpile stewardship, Phys. Today 53(12), 44–50.
Krass, A (1997). The United States and Arms Control, Praeger, New York.
Krepon, M. and D. Caldwell (1991). The Politics of Arms Control Treaty Ratification, St. Martin’s
Press, New York.
Krepon, M. and M. Umberger (1988). Verification and Compliance, Ballinger, Westport, CT.
Moynihan, M. (2000) The scientific community and intelligence collection, Phys. Today 53(12),
51–56.
National Academy of Sciences (2002). Technical Issues Related to the Comprehensive Nuclear Test
Ban Treaty, National Academy Press, Washington, DC.
Nelson, R. (2002). Low-yield earth-penetrating nuclear weapons, Sci. Global Secur. 10, 1–20.
Office of Technology Assessment (1991). Verification Technologies, OTA, Washington, DC.
Richelson, J. (1998). Scientists in black, Sci. Am. 278(2), 48–55.
Sabbins, F. (2000). Remote Sensing, Freeman, San Francisco, CA.
SDI Technology (1988). Survivability and Software, Office of Technology Assessment.
Sykes, L. (1996). Dealing with decoupled nuclear explosions under a Comprehensive Test
Ban Treaty, in Monitoring a Comprehensive Test Ban Treaty, E. Husebye and A. Dainty (Eds.),
Kluwer, Amsterdam, the Netherlands.
Tsipis, K., D. Hafemeister and P. Janeway (Eds.) (1986). Arms Control Verification, Pergamon,
Washington, DC.
US Arms Control and Disarmament Agency (1996). Arms Control and Disarmament Agree-
ments, ACDA, Washington, DC.
von Hippel, F. and R. Sagdeev (Eds.) (1990). Reversing the Arms Race, Gordon and Breach,
New York.
P1: GFZ
5
Nuclear Proliferation
We are here to make a choice between the quick and the dead. That is our
business. Behind the black portent of the new atomic age lies a hope which,
seized upon with faith, can work our salvation. If we fail, then we have damned
every man to be the slave of fear. Let us not deceive ourselves: We must elect
world peace or world destruction.
“Science has torn from nature a secret so vast in its potentialities that our
minds cower from the terror it creates. Yet terror is not enough to inhibit the
use of the atomic bomb. The terror created by weapons has never stopped man
from employing them. For each new weapon a defense has been produced, in
time. But now we face a condition in which adequate defense does not exist.”
[Bernard Baruch, US Representative to UN Atomic Energy Commission,
June 14, 1946]
5.1 Proliferation: Baruch to Iraq

In a dramatic moment before the United Nations, Bernard Baruch described the
American plan to internationalize and control the atom. He described in biblical
fashion the choice between “the quick from the dead” (taken from the Apostle’s
Creed) resulting from the global spread of nuclear weapons. This prediction came to
fruition 15 years later when presidential candidate John F. Kennedy gave a warning
in the third debate with Vice President Richard Nixon:
“There are indications because of new inventions, that 10, 15, or 20 nations will
have a nuclear capacity, including Red China, by the end of the presidential
office in 1964. This is extremely serious. There have been many wars in the
history of mankind and to take a chance now and not make every effort that we
could make to provide for some control over these weapons, I think, would be
a great mistake.”
[John Kennedy, in the presidential debates with Richard Nixon,
October 13, 1960]
Kennedy’s projection of 20 nuclear weapon states was correct, but it took a few
more decades to arrive at eight consisting of United States, Soviet Union, United
105
P1: GFZ
106 5. Nuclear Proliferation
Table 5.1. Nuclear Weapon State status. Twenty-five nations are categorized in terms of
their progress to nuclear weapons with the dates of the first (fission/fusion) tests listed for
the five nuclear weapon states (NWS as defined in the NPT) and the two emerged NWS.
Dates for the other states indicate the years of their active nuclear programs.
5 NWS: US (fission 1945/fusion 1952), FSU/Russia (1949/1953), UK (1952/1957), France (1960/1966),
China (1964/1967).
4 defacto NWS: India (1974/1998-claimed), Pakistan (1998), Israel (1979?), North Korea (2006)
4 former defacto NWS: South Africa (1979–1993), Ukraine, Belarus, Kazakshtan (1991–94).
10 former nuclear weapon programs
3 recent nuclear weapon programs: Iran (enrichment), Iraq (1975–1991), Libya (1990–2004).
Kingdom, France, China, India, Israel, Pakistan, and North Korea for a total of nine.
South Africa gave up its six nuclear weapons in 1993. If one counts the former Soviet
republics of Belarus, Kazakhstan, and Ukraine, which gave up their Soviet weapons
in the 1990s, the total would be 13. This exceeds Kennedy’s lower bound of 15 when
the striving nations of Iraq, Libya, and Iran are included. The total information in
Table 5.1 exceeds Kennedy’s upper bound of 25.
Commercialization of nuclear power raises three issues: (1) Nuclear proliferation
from the spread of nuclear technology and materials to nations that in turn develop
nuclear weapons; (2) nuclear safety resulting from the release of large amounts
of radioactivity from reactor fuel (including spent fuel fires); (3) the disposal of
nuclear waste to underground or surface storage sites. This chapter deals only
with the proliferation issue, as safety and wastes are covered in Chapter 7. In the
author’s view, the severity of these issues is ranked as follows: Proliferation is of
more concern than reactor safety, which is of greater concern than waste disposal.
5.1.1 Connection Between Peaceful and Military Atoms?

Does the development of commercial nuclear power contribute to the spread of
nuclear weapons? Proliferation policy is more complicated than superpower strate-
gic weapons policy, since nuclear power supplies commercial energy to 30 nations,
a number that could grow to 50. The START/SORT and ABM treaties involve only
two nations (United States and Russia) and they do not impact commercial energy
supplies. In 2000, 30 nations produced 16% of the world’s electricity with nuclear
power for a total capacity of 350 GWe . The United States had 98 GWe nuclear
power in 2000, producing 20% of its electricity. The US load factor rose to 85%,
as the new fuels remain longer in reactors, giving less frequent shutdowns. Many
nations depend heavily on nuclear energy; these include France (63 GWe , 77% nu-
clear), Japan (44 GWe , 34%), Russia (20 GWe , 15%), Germany (20 GWe , 31%), South
Korea (13 GWe , 42%), United Kingdom (13 GWe , 23%), and Belgium (5 GWe , 55%).
Some of these states have their spent fuel reprocessed (France, UK, Japan, Belgium)
in France, UK, and Russia, but the foreign contracts are coming to an end. Japan
has almost completed its large reprocessing plant. Others nations are storing spent
fuel, waiting for a geological repository.
P1: GFZ
5.1. Proliferation: Baruch to Iraq 107
Nations obtain nuclear weapons more out of mutual mistrust with a neighboring
state than fear of a distant superpower. Such a nearest neighbor interaction is ex-
emplified by the (first/second) nation duos: US/USSR, USSR/China, China/India,
India/Pakistan, North Korea/South Korea, Israel/Arab states, Argentina/Brazil,
and China/Taiwan. Proliferation started with Klaus Fuchs and Ted Hall passing
American nuclear secrets to help Soviet designers. Friendly cooperation in the US
Manhattan Project helped Britain on its way, and it has been stated that the United
States gave indirect assistance to the French. The Soviets gave major assistance to
China to help it become a nuclear weapons state. The French helped Israel with the
sale of the Dimona reactor and associated reprocessing technology. Canada assisted
India with the sale of the Cirus reactor, to which the United States supplied some
heavy water and some reprocessing technology. China helped Pakistan with de-
signs, materials, and missiles. Pakistan’s A.Q. Kahn sold centrifuges & weapon de-
signs to North Korea, Iran and Libya. And so the story goes. Such events prompted
Tom Lehrer to write his song, “Who’s Next”
First we got the Bomb, and that was good,

Cause we love Peace and motherhood.
Then Russia got the bomb, but that’s OK.
The balance of power’s maintained that way.
Who’s next? [France, Egypt, Israel]
.........
Luxembourg is next to go.
Then who knows, maybe Monaco.
We’ll try to stay serene and calm
When Alabama gets the Bomb.
Who’s Next?
[Reprise Records, 1965]
Nuclear weapon programs usually grow out of dedicated national programs

(that is, programs dedicated to nuclear weapon development) rather than civil
nuclear power programs. But there often is a connection between the peaceful
and military fuel that cycles through reactors that make plutonium, or centrifuges
that make highly enriched uranium (HEU). India, Pakistan, Iraq, Iran, and other
countries used research power programs as a cover to hide military programs.
Proliferation watchers are rightly concerned when nations with small electrical
power grids move toward nuclear power. If a nation risks destabilizing its small
grid by connecting it to large nuclear power stations, there could be some other
reason. Similarly, small nations that plan to build an enrichment or reprocessing
plant could be establishing a weapons program since it is not cost effective to enrich
or reprocess on a small scale. As we discuss later, reactor-grade plutonium (RgPu)
from civil power plants in aspiring nuclear states could be used for weapons, but the
Pu weapons in those states would be of a lesser quality. Still, reactor-grade Pu can be
made into weapons, and that is why its production and storage is monitored by the
International Atomic Energy Agency (IAEA). Thus far, the eight nuclear weapon
P1: GFZ
Figure 5.1. Iraq’s supergun. The

350-mm diameter “supergun”
that Iraq tested at Jabal Hamrayn,
200 km north of Baghdad to shoot
at Israel. [United Nations Inspec-
tion Team, 1991]
states with plutonium weapons have only used weapons-grade plutonium (WgPu)
from dedicated reactors. See Figs. 5.1 & 5.2.
5.1.2 Atoms for Peace

President Dwight D. Eisenhower gave his “Atoms for Peace” speech to the United
Nations (December 8, 1953), when he proposed that an IAEA should impound,
store, and protect fissile materials. Eisenhower claimed that nuclear fuel could
be “proliferation-proof”: “The ingenuity of our scientists will provide special safe
conditions under which such a bank of fissionable material can be made essentially
immune to surprise seizure.” Did Eisenhower mean plutonium would be immune
to surprise attack because of excellent physical protection and safeguards? Or did he
mean that ample amounts of the spontaneous neutron-emitter 240 Pu would prevent
the use of plutonium in bombs? At that time, scientists incorrectly believed that
20% 240 Pu gave enough early neutrons to preinitiate a chain reaction and greatly
reduce the yield for all weapons designs. This belief has been proven false.
P1: GFZ
Figure 5.2. Iraq’s EMIS enrichment plant. View of control room under construction at the
Tarmiyah Industrial Enrichment Plant for uranium enrichment with electromagnetic isotope
separation (EMIS). [UN, 1991]
An interesting question to debate is whether actions subsequent to the Atoms

for Peace speech reduced or enhanced global proliferation of nuclear weapons?
There is no simple answer, but this author takes a long-range point of view. Eisen-
hower did the right thing because his policies fostered the establishment of the
IAEA in 1957 and the Nuclear Nonproliferation Treaty in 1970. Alone, the United
States has never had the leverage to dictate its wishes to all nations. With all na-
tions agreeing on a common framework of rules and inspections, the world can
act in concert. Without the IAEA and NPT, which form an international nuclear
regime, the world community would not have an effective way to exercise con-
trol over NNWS weapons programs and plutonium/uranium stockpiles. With the
passage of time, technical barriers to making nuclear weapons became smaller as
secrets leak and technologies improve. Without a global, political commitment to
nonproliferation, the march of technology would be unstoppable.
On the other hand, it is clear that proliferation was accelerated under Atoms
for Peace, as plutonium reprocessing was declassified and taught to foreign scien-
tists at Argonne National Laboratory. The Indian scientists who reprocessed pluto-
nium at Argonne went home to help India produce its bombs. The NPT also gave
cover to weapons programs that might not have existed otherwise. The Iraqi and
North Korea violations took place, however, at undeclared locations that were not
inspected. If IAEA members had allowed the organization to act more vigorously,
it would take stronger actions, such as challenge inspections. South Africa devel-
oped its gaseous nozzles and nuclear weapons essentially, but not entirely, alone.
P1: GFZ
Table 5.2. World military inventories of HEU and

plutonium. These amounts for 2001 do not include 200 tons
of separated civilian plutonium owned by Britain (60 ton),
France (40 ton), Germany (25 ton), Japan (30 ton), and Russia
(30 ton). There are also 1000 tons plutonium in civil spent
fuel rods and 20 tons of civil HEU. [D. Albright, Institute for
Science and International Security, www.isis-online.org]
HEU (tons) Pu (tons) Warheads
Russia 970 130 20,000
US 635 100 10,000
France 24 5 450
China 20 4 400
UK 15 7.6 185
Israel – 0.51 100
India – 0.31 45–95
Pakistan 0.7 0.005 30–50
North Korea – 0.035 –
South Africa 0.4 – –
Total 1665 248 31,200
They succeeded with an expenditure of hundreds of millions of dollars, much less

than what Saddam Hussein spent on his covert unsuccessful program. There is no
doubt about it, openness on nuclear matters contributed to proliferation.
The world had 438 nuclear reactors with 351 GWe of power in 2000 and 33 addi-
tional reactors under construction. The IAEA mission of promoting nuclear power
and control of nuclear materials has grown substantially as its membership has
grown to 140 states (Table 5.2). The 2000 safeguards budget of $80 million covered
902 facilities containing 642.8 tons of plutonium in spent fuel rods, 72.2 tons sepa-
rated Pu, 10.7 tons in mixed oxide (MOX) rods in reactors, and 21.8 tons of HEU.
The 10,264 person-days of inspections on 2467 inspections applied 25,484 encrypted
seals to containers, reviewed 5226 videotapes and analyzed 626 samples. The IAEA
appears to have a good record on declared sites with no major mistakes, as the fa-
mous Iraqi and North Korea violations involved undeclared locations. After the Gulf
War the IAEA changed its procedures (1) to increase the use of intelligence informa-
tion from large nations, (2) to take environmental samples in search of clandestine
enrichment and reprocessing, and (3) to establish special inspection procedures for
undeclared sites. It is not clear how effective these additional measures will turn
out to be since they require more funding, but it is clear that a multilateral approach
is essential for managing proliferation in a world of 185 nations. The 1977 Office of
Technology Assessment report on proliferation concluded the following:
“In the long run two general rules apply: (a) Solutions to the proliferation prob-
lem will have to be found primarily, though not exclusively, through multilateral
actions, and (b) the extent of US influence will vary from country to country.”
[Nuclear Proliferation and Safeguards, OTA, Washington, DC, 1977]
P1: GFZ
5.1.3 The NPT

The Nuclear Nonproliferation Treaty created a discriminatory regime with two
classes of nations. The have nations, the “big five” of World War II, which are defined
in the NPT as nuclear weapon states (NWSs: United States, Russia, UK, France,
and China) that obtained both UN veto power and nuclear weapons. The have-not
nations (NNWSs, nonnuclear weapon states) became the remaining 180 parties to
the NPT. The main NPT holdouts are India, Israel, and Pakistan, while Iraq and
North Korea are NPT parties that are “not in good standing.”
There is a disparity in obligations as set down by the treaty. The NPT requires
safeguard inspections on NNWS nuclear facilities, but not on NWS facilities. Indeed,
some of the NWS have volunteered their facilities for inspection, but the IAEA does
not have the funds to carry this out. The NPT strongly encourages the NWS to assist
with nuclear power and research programs for the NNWS, a requirement that has
been interpreted to include the use of plutonium and weapons-grade uranium fuel.
Such cooperation has taken place, but the Carter administration constrained this
use of plutonium and HEU, an act that was condemned abroad, but which now
has gained momentum.
The 1970 NPT calls upon the NWS to end the nuclear race in Article VI:
“Each of the Parties to the Treaty undertakes to pursue negotiations in good faith on effective
measures relating to cessation of the nuclear arms race at an early date and to nuclear
disarmament on a Treaty on general and complete disarmament under strict and effective
international control.”
The 1994 START and the 2002 SORT treaties are a step in that direction, but
nonnuclear weapon states want further progress. But, the 1999 defeat of the Com-
prehensive Test Ban Treaty in the Senate was a step back from the US promise at the
time of the 1995 indefinite extension of the NPT to adopt a permanent test ban. The
CTBT is regarded by NPT nonnuclear states as the litmus test on NWS intentions.
The defeat of the CTBT by the US Senate, the Indian–Pakistani nuclear tests of
1998, the Iraqi, North Korean, and Iranian nuclear programs, the modest progress
on strategic offense weapons and the demise of the ABM treaty are dangerous
indicators of problems with the NPT compact that should not be ignored.
5.1.4 Nonproliferation Policy

A complete discussion of nonproliferation contains many elements beyond over-
sight by the NPT and IAEA. Its points include consideration of the following:
r Plutonium economy. The United States abandoned plutonium recycling with the
circa 1980 cancellation of the breeder reactor and reprocessing. This course has
been adopted by many other nations.
r Reduction in excess weapons-grade materials. The United States and Russia agreed
to each dispose of 34 tons of WgPu, as well as 500 tons of Russian and 174 tons
of US HEU.
P1: GFZ
r An assured supply of uranium. A large way to reduce interest in recycling Pu in

reactor fuel demand is for the United States to be an assured supplier of less
costly, LEU. Spent fuel contains 1% Pu. It is for this reason that the United States
maintains control over reprocessing rights on spent fuel that has been enriched in
the US or irradiated in US supplied reactors. Uranium supplies have been helped
by a dramatically slowed growth in nuclear power, increased uranium efficiency
of reactors, and conversion of weapons-grade uranium of the cold war era into
reactor fuel.
r Nuclear suppliers group. The supplier nations agreed to constrain exports of sensi-
tive nuclear fuel facilities (enrichment and reprocessing) and require that all the
importer’s nuclear facilities be under safeguards.
r Export Criteria. The Indian bomb of 1974 created the climate for the passage of the
Nuclear Nonproliferation Act of 1977, which requires safeguards on all nuclear
facilities before exports are allowed. It also established criteria for the United
States when considering if it would allow reprocessing of US-origin spent fuel
by other nations.
r Sanctions. When an NNWS moves toward the bomb, it will be denied nuclear
power exports, military equipment, and other items of commerce. These sanc-
tions were mostly removed 2 years after the Indian and Pakistani 1998 nuclear
tests.
r Russian weapons usable materials. It is imperative that Russian warheads and
weapons-grade nuclear materials remain under firm Russian accounting and
control. The US Cooperative Threat Reduction programs have significantly im-
proved this situation, but much more needs to be done.
r Spent fuel storage. There are 32,000 tons of US-origin spent fuel at storage sites
around the world. The United States does not want to accept these materials nor
does it want the spent fuel to be reprocessed. Russia passed a law in 2001 that
would allow it to establish a large storage site, which could be a useful constraint
on the plutonium in the event that it is not reprocessed.
r Preemption. Case by case?
5.1.5 Special Nuclear Material

The IAEA unit for measuring weapons-grade nuclear material is called a significant
quantity. It is defined as 25 kg of 235 U in HEU or 8 kg of Pu. Nuclear weapons can
be made with less than a significant quantity by reflecting neutrons back into the
warhead, by compressing fissile metal to higher densities with explosives, and by
fission boosting with tritium and deuterium. Uranium is enriched in 235 U content
by using the laws of physics, while plutonium is separated from spent fuel rods
in reprocessing plants by using the laws of chemistry. Nations often began nuclear
weapons production with uranium weapons and moved up to plutonium weapons;
however Russia, France, India, Israel, and North Korea began with plutonium.
See Figs. 5.3 and 5.4 for plutonium gamma-ray spectra and Table 5.3 for nuclear
production facilities.
P1: GFZ
SVNY342-Hafemeister
March 27, 2007
15:32
Figure 5.3. Pu300 /Pu600 /Pu900 monitoring. By monitoring gamma-ray windows near 300 keV, 600 keV, and 900 keV it is possible to
determine (1) Pu presence, (2) Pu age since reprocessing, (3) Pu content to determine if Pu is weapons-grade, and (4) absence of
plutonium oxide (along with other measurements). In addition, gamma-ray spectra can give a minimum Pu mass estimate. However, an
estimate for Pu mass is more accurately obtained through neutron counting. [Technology R&D for Arms Control, Department of Energy,
2001]
5.1. Proliferation: Baruch to Iraq
113
P1: GFZ
Figure 5.4. Pu quality. The gamma-ray window between 635 and 665 keV displays transi-
tions from both 239 Pu and 240 Pu. Weapons-grade Pu contains 6% 240 Pu, while reactor-grade
contains more than 20% 240 Pu. Monitoring at the Mayak storage facility near Ozersk will use
a ratio of 0.1 = 240 Pu/239 Pu to separate the two materials. [Technology R&D for Arms Control,
Department of Energy, 2001]
5.1.6 September 11, 2001

The terrorist attacks on the World Trade Center and the Pentagon killed 3000 and
caused a paradigm shift from fear of nations to fear of terrorists. It is often ar-
gued that subnational terrorists will more likely use conventional explosives and
biochemical weapons than nuclear weapons. It is far easier to use trucks full of
fertilizer, divert airplanes, or drop aerosol bombs in subways than it is to obtain
nuclear materials and assemble a viable nuclear weapon. But reports of an attempt
by al Qaeda to obtain nuclear materials gives one pause. Terrorists could steal
Pu/HEU to sell to proliferant nations. Terrorism involves many possible paths and
responses. How can the world really know what is in 15 million shipping canisters
on the high seas each day? Pu/HEU can be identified at airports, but only reliably
with high-tech, pulsed neutron sources. We will not find solutions to global ter-
rorism only in improved detection technologies, but rather we must also consider
broader political accommodations with the have-not nations.
Table 5.3. Enrichment and reprocessing production plants. These national facilities are
operating or under construction. The list does not include facilities that have been closed.
[D. Albright, 1997]
Argentina (enrichment 1/reprocessing 0), Belgium (0/1), Brazil (1/0), China (2/1), France (1/1),
Germany (1/0), India (1/3), Israel (0/1), Japan (3/2), the Netherlands (1/0), Pakistan (1/1), Russia
(4/3), South Africa (1/0), UK (2/2), US (1/0).
P1: GFZ
The 110-story World Trade Center was designed to withstand a Boeing 707, but
it could not withstand prolonged (1–2 h) 2000◦ F temperature from full tanks of
burning jet fuel. Building 7 of the World Trade Center with 47 stories also collapsed,
though it was not hit by a jet plane. It now appears that this collapse resulted from
the ignition of 1000 bbl of diesel fuel that was stored in the basement. Terrorism
is considered in problems on collapsing buildings (problem 5.20) and destroying
anthrax in letters (problem 5.21). Chapter 7 calculates the radioactive plumes from
reactor and plutonium-weapon accidents, which has similarities to “dirty bomb”
dispersal of radionuclides with conventional explosives.
5.1.7 Preemptive Counterproliferation

The multilateral NPT/IAEA regime is not perfect, but it is all we have. Or is it? The
June 1981 Israeli destruction of Iraq’s Osirak reactor was decried by most mem-
bers of the UN and the IAEA since Iraq was a member of these organizations “in
good standing.” Yet, today there is little hand wringing over the unilateral attack on
Osirak, since it could have produced plutonium for Iraqi weapons. This preemptive
attack rolled back proliferation (counterproliferation), and such attacks may become
more likely. As with all policies, there are upsides and downsides. Preemptive roll-
backs get the job done quickly, while international diplomats argue and postpone.
But preemptive attacks shake the foundations of international processes and due
process, unless international organizations are part of the process. An examina-
tion of motives for preemptive attacks points to inconsistencies, but these have to
be balanced against the degree of threat from proliferation. The military has long
defined threats in terms of both capabilities and intentions. Preemptive counterpro-
liferation appears to be driven more by the perceptions of the intentions of nations,
since many states certainly have capabilities. The 2002 US National Security Strat-
egy states it clearly: “We will not hesitate to act alone, if necessary, to exercise
our right of self-defense by acting preemptively.” But in soured relationships it
is often difficult to separate worst-case interpretations from lesser interpretations.
Ultimately, counterproliferation can lead to assassinations of leaders. This outcome
is constrained by Presidential Executive Order 12333, which, nonetheless, can be
quickly nullified without the approval of congress. One might ask if less stringent
criteria for assassinations could rebound to put leaders in harm’s way. As the third
millennium begins, there are ominous signs from North Korea, which withdrew
from NPT (January 11, 2002) and from Iran, which was charged by the IAEA with
violating the NPT for building clandestine centrifuges (June 6, 2003). How will
these events be handled? These issues are part of the continuing national debate.
The January 2005 elections in Iraq were encouraging, but will the new Iraq be
stable?
5.1.8 Biological and Chemical Weapons

Both the Chemical Weapons Convention (CWC) and Biological Weapons Conven-
tion (BWC) ban the production, acquisition, stockpiling, transfer, and use of these
P1: GFZ
materials. However, the CWC has a verification protocol, while the BWC does
not. After the anthrax attack in the US mail, there has been heightened concern
about BW and CW materials. In general, chemical weapons are considered about
as lethal per unit mass as conventional explosives and much less lethal than nu-
clear weapons. However, biological weapons could be more lethal than nuclear
weapons if they are dispersed effectively and widely. Experts generally agree that
BW production is not technically difficult, but effective dissemination is much more
difficult. A dose of 10,000 anthrax spores can be lethal, while plague needs only 100
to 500 organisms and smallpox needs even less at 10 to 50 organisms, which can
spread to others. Sophisticated devices, such as the Handheld Advanced Nucleic
Acid Analyzer, can detect pathogens in the field by examining DNA of samples
and comparing to known DNA sequences of various pathogens. Unfortunately,
current research exacerbated matters with development of mutant anthrax that is
resistant to antibiotics, and mouse pox that could circumvent the lack of smallpox
samples.
5.2 Uranium Enrichment

Even if plutonium were effectively controlled, HEU (at 90% 235 U) offers another
path to nuclear weapons production. For one thing, it is easier to make HEU gun
weapons as compared to Pu implosion weapons. Secondly, it was thought that
HEU was harder to obtain than plutonium, but the enrichment barrier to HEU
has lowered over the years. This was proven when South Africa successfully used
gaseous nozzles to obtain HEU for its six nuclear weapons. On the other hand,
Iraq failed with its electromagnetic isotope separation (EMIS), but Pakistan has
succeeded with gas centrifuges (and Iran is starting). This section estimates prop-
erties of some enrichment technologies used for obtaining the various uranium
categories listed in Table 5.4.
5.2.1 Gaseous Diffusion

Lighter 235 UF6 molecules go faster than molecules containing 238 U, allowing 235 U to
pass through small pores in a membrane at higher rates. The diffusion barrier used
in gaseous diffusion separation of 235 U is made of sintered, 1-μ diameter nickel
Table 5.4. Enriched uranium isotopic composition

234 U 235 U 238 U
Weapons-grade U 1% 93.3% 5.5%

HEU < 1% > 20% < 80%
Natural U 0.0054 0.711% 99.3%
Depleted U – 0.2% 99.8%
P1: GFZ
5.2. Uranium Enrichment 117
spheres with 25-nm pore diameters. The ratio of 235 UF6 to 238 UF6 velocity is
v235 /v238 = [m238 /m235 ]1/2 = [(238 + 6 × 19)/(235 + 6 × 19)]1/2 = [352/349]1/2
= 1.0043, (5.1)
which gives a gain of 0.4% per stage. It takes 1000 gaseous diffusion stages to
make 3.2% enriched fuel and 3500 stages to make 90% enriched weapons uranium.
Gaseous diffusion is more effective for light molecules and less effective for the
higher-mass UF6 molecules because diffusion rates depend on the mass ratio. This
is in contrast to gaseous centrifuge and gravitational separation, which depend on
mass differences.
5.2.2 Gravitational Separation

The United States Manhattan Project used gravitational separation of isotopes to
produce slightly enriched feedstock for gaseous diffusion separators, which then
fed electromagnetic separators. Gravitational separation combines a differential
gravitational force with thermal mixing to increase the ratio RU = 235 U/238 U at a
height h above the bottom of the tube. Employment of Boltzman statistics gives an
enrichment factor
RU (h)/RU (0) = exp[−(m1 − m2 )gh/Rgas T], (5.2)
where mass m is in kilogram-moles, g is 9.8 m/s2 , and the universal gas constant
Rgas is 8.3 J/K. For the case of a 10-m tube height h at 330 K, 235 U is enriched by a
factor

−(0.349 − 0.352)(10)(g)
RU (10)/RU (0) = exp = 1.0001, (5.3)
(8.3)(330)
where g is 9.8 m/s2 , the 235 UF6 kg molar mass is 0.349 kg and 238 UF6 is 0.352 kg.
The gravitational-thermal enrichment of 0.01% per stage is much less than gaseous
diffusion’s 0.43% per stage. Note that the enrichment factor depends on the mass
difference, not the mass ratio.
5.2.3 Gaseous Centrifuges

In 1940, 2 years before the Manhattan Project, Germany began centrifuge ex-
periments with force fields greatly exceeding gravity. Centrifuges can produce
considerable enrichment per stage, but the German centrifuges were destroyed at
the high rotational velocities. Heavier molecules diffuse close to the inside surface
of the centrifuge tube while lighter ones diffuse toward the center of the tube. At
ultrahigh speeds the two components of the gas are in a thin layer near the tube
wall. A current is formed with lighter gases rising to the top and heavier gases
falling to the bottom, where they are collected. Today’s centrifuges use carbon or
P1: GFZ
glass fiber tubes, since the tensile strengths of aluminum and other metals are too
small for the rotational stress. The centrifuge enrichment factor is
RU (r ω)/RU (0) = exp[(m1 − m2 )(ω2r 2 )/2Rgas T]

= exp[(0.352 − 0.349)(50002 × 0.12 )/(2)(8.3)(330)] = 1.15, (5.4)
with r = 0.1 m and ω = 5000 rad/s ( f = 800 rev/s). The centrifuge gain of 15% per
stage is 30 times larger than that of a diffusion stage. Centrifuges need only a dozen
stages to obtain 3.2% reactor fuel compared to a thousand stages for diffusion. A
smaller centrifuge plant could be capable of producing 25-kg weapons-grade ura-
nium in 2 months. Such a plant can be built clandestinely in a space of 60 m by 60 m
and need only tens of MWe as compared to one or more GWe for gaseous diffusion
plants. The equilibrium time for a centrifuge plant is only minutes, allowing plant
operators to shift LEU production piping to HEU-production piping, but still with
difficulty. To make sure this is not happening, the IAEA Hexapartite agreement
allows for inspections with 2-h notice to intrusively monitor isotope ratios and
sealed valves and to perform some remote monitoring of certain pipes.
5.2.4 Electromagnetic Isotope Separators

Electromagnetic isotope separators (EMIS) were developed by Ernest Lawrence to
produce HEU at Oak Ridge. These were called “Calutrons” since Lawrence was
from the University of California. They produced enough HEU for the Hiroshima
weapon, but they used slightly enriched feed from gravitational and gaseous diffu-
sion. Iraq surprised the world by choosing an improved EMIS technology, but it did
not complete the construction before their EMIS were discovered and destroyed by
UN inspectors. Iraq was building 140 EMIS separators using 300 MWe to produce
30 kg/year of weapons-grade uranium. The radius of an ion beam in a uniform
magnetic field is
r = mv/qB, (5.5)
where m is mass, v is velocity, q is ion charge, and B is magnetic field. The ion’s
energy comes from the electrical potential, qV = 1/2 mv2 , giving an ion radius of
r = (2Vm/q )1/2 /B. (5.6)
The Calutron radius is 1.2 m using B = 0.34 T, V = 35, 000 volts, q = 1.6 × 10−19
coulombs, and m = 3.9 × 10−25 kg. The fractional change in the radius between
235
U and 238 U ions is
r/r = (m/m)/2 = (3/238)/2 = 0.0063 = 0.63%, (5.7)
giving a separation of r = (0.0063)(1.2 m) = 8 mm, sufficient to obtain HEU in

two stages. Calutrons were quickly superceded by gaseous diffusion in the United
States since considerable feedstock is lost during ionization.
P1: GFZ
5.3. Separative Work Units 119
5.2.5 Laser Isotope Separation

The slightly smaller volume of 235 U nuclei provides enough different electrostatic
interaction with s-electrons to separate 235 U from 238 U. Multiple excitations with
tunable dye lasers and other lasers ionize 235 U in atomic vapor or in UF6 molecules,
allowing electric fields to separate the 235 U ions from 238 U atoms or molecules.
Laser isotope separation (LIS) could give easier access to weapons-usable HEU,
thus making it possible to avoid using nuclear reactors to produce plutonium.
Making high enrichments from LIS is complicated by the charge-exchange reaction
involved, namely
235
U+ + 238 U = 235 U + 238 U+ , (5.8)
238 235
which adds U ions to the U ion stream, diminishing enrichment levels. Prob-
lem 5.10 deals with estimating energy differences between 235 U and 238 U nuclei.
5.2.6 Aerodynamic Nozzles and Helicons

UF6 molecules turn very tight corners in a Becker nozzle; the heavier 238 UF6 is
deflected less than 235 UF6 because of its larger inertial mass. Isotopes are separated
with a knifeblade into two streams after they turn a corner. Gas in a radial turn of
0.1 mm at 400 m/s experiences tremendous centripetal acceleration,
a c = v2 /r = (400 m/s)2 /(10−4 m) = 1.6 × 109 m/s2 = 160 million g. (5.9)
South Africa chose a similar approach, but instead projected UF6 at right angles
in a tightening cone, with the spiraling 238 UF6 revolving more to the outside and
235
UF6 revolving more to the inside. This process is similar to a centrifuge with
a nonrotating casing. Again, million’s of g’s are developed in a tight geometry.
South Africa’s weapons program was amazing in that it was carried out mostly in
isolation at a relatively low cost of hundreds of millions of dollars.
5.2.7 Chemical Ion Exchange

France and Japan developed pilot plants that took advantage of small isotopic mass-
dependent differences in chemical reaction rates. Using catalysts, they observed
that lighter isotopes tend to preferentially bind to more loosely bound compounds,
while heavier isotopes tend to bind to the more strongly bound compounds. Mixing
the two compounds causes 235 U to flow from the more tightly bound compound
to the loosely bound compound. The two compounds are separated chemically,
giving an enriched product.
5.3 Separative Work Units

This section is intended for the more dedicated readers, as the topic of separa-
tive work units (SWUs) is cumbersome. Separation of isotopes cannot be 100%
P1: GFZ
complete, except in very small samples. The enrichment process changes the iso-
topic ratio of the feedstock by increasing the ratio of the desired isotope in the
product and decreasing its ratio in the waste (the tails). Mixing separated isotopes
increases chaos, raising the system’s entropy, and, conversely, isotope separation
creates order and lowers entropy. Separation of isotopes lowers the entropy of feed
(F ) to the sum of the entropies of product (P) and waste (W). The value V of a mixture
is closely related to the statistical mechanics definition of entropy, S = n ln(n), where
n is the number microstates. The thermodynamic entropic change (S = Q/T)
is proportional to energy consumption for a given technology at constant tempera-
ture T. The SWU is the difference in the values V needed to convert feed to product
plus waste (usually called tails). The fractional isotopic abundances of 235 U is f ,
between 0 and 1. The value is given as a function of f for feed ( f F ), product ( f P ), or
waste ( f W ):

f
V( f ) = (2 f − 1) ln . (5.10)
1− f
Separative work is the difference between the values of the output and input.
The total separative value of a sample is its value times its mass, hence its units in
kg SWU or tonne-SWU. Separative work done is the difference of the total values,
a calculation that gives the number of SWUs needed to do the separation:
number of SWUs = P V( f P ) + WV( f W ) − F V( f F ) (5.11)
where F = P( f P − f W )/( f F − f W ) and W = P( f P − f F )/( f F − f W ). Enrichment

plant sizes are given in units of tonne-SWU/year. If the mass of the product
P in Eq. 5.11 is 1 tonne, the number of SWUs is in units of tonne-SWU. Since
it takes a fixed amount of energy to produce a SWU, the product of ton SWU
is essentially energy, similar to mgh for lifting mass in a gravitational field.
Three situations for a plant with a capacity of 1000 tonne SWU/year are as
follows:
5.3.1 3.2% Fuel

Reactors of the 1970s used 3.2% enriched uranium fuel, obtained from 0.72% natural
feed with 0.2% tails. By inserting f P = 0.032, f F = 0.0072, and f W = 0.002 into the
above formulas, we obtain
number of SWUs = 4.7P. (5.12)
This means it takes 4.7 kg SWU to obtain 1 kg of 3.2% enriched fuel from natural
uranium with tails of 0.2%. It takes F = 5.8 kg of natural uranium feed to make 1
kg of 3.2% fuel. A 1000-tonne SWU/year plant could produce P = 210 tonne/year
(1000/4.7) of 3.2% product from a feed F = 1230 tonne/year, while rejecting 0.2%
tails at W = 1020 tonne/year.
P1: GFZ
5.4. Nonproliferation in the Former USSR 121
5.3.2 4.4% Fuel

Since the 1970s, the time between shutdowns for the purpose of refueling increased
from 12 to 18 months by the use of higher enriched fuel of 4.4% 235 U. This change
helped increase nuclear power plant load factors to 90%, making plants more prof-
itable, while using less uranium and producing less plutonium and less spent fuel.
The ratio of enrichments (4.4% versus 3.2%) is roughly the ratio of times between
shutdowns. New fuels require better materials that remain viable for a longer time
in the reactor. Using the above equations, it takes 8.2 tonne of U and 7 tonne SWU
to make 1 tonne of 4.4% fuel from natural U with 0.2% tails. The cost of uranium
and separative work to make 1 kg of 4.4% fuel is about
Ckg 4.4% = (11 kg U)($30/kg U) + (7 kg SWU/kg)($90/kg SWU)

= $330 + $630 = $1000. (5.13)
5.3.3 90% HEU

It takes 226 kg SWU and 174 kg of natural uranium to make 1 kg of 90% HEU
with 0.2% tails. A 1000 tonne SWU/year plant can make 4.4 tonne/year of HEU
from 765 tonne/year of natural uranium, rejecting 761 tonne/year of 0.2% tails.
The approximate cost of uranium and separative work to make 1 kg of HEU is
about
Ckg HEU = (174 kg U)($30/kg U) + (225 kg SWU)($90/kg SWU)

= $5000 + $20,000 = $25,000. (5.14)
The HEU in a warhead is worth about $500,000 (20 kg × $25,000/kg). Note that
about 50% of the seperative work for HEU is spent to make the LEU precursor
(Problem 5.13).
5.4 Nonproliferation in the Former USSR

The CIA’s 2002 Annual Report to Congress on the Safety and Security of Russian Nuclear
Facilities and Military Forces contained the following comments on Russian nuclear
materials:
r In 1992, 1.5 kg of 80% enriched weapons-grade uranium were stolen from the
Luch Production Association.
r In 1994, 3.0 kg of 90% enriched weapons-grade uranium were stolen in Moscow.
r Although not independently confirmed, reports of a theft in 1998 from an un-
named enterprise in Chelyabinsk Oblast are of concern according to Viktor
Yerastov, chief of Minatom’s Nuclear Materials Accounting and Control Depart-
ment: The amount stolen was “quite sufficient material to produce an atomic
bomb,” the only nuclear theft that has been so described.
P1: GFZ
5.4.1 Pu and HEU Stockpiles

Let us approximate the size of Russia’s stockpile of plutonium and weapons-grade
uranium. In spite of the crudeness of the calculations we will gain understanding for
their weapons-usable materials. The Soviets began the process by placing 410,000
tonnes of natural uranium fuel in plutonium production reactors, which was then
reprocessed as spent fuel to obtain plutonium. Residency of a few months in reac-
tors lowered the 235 U content from 0.711% to 0.667%, as given by Oleg Bukharin
of Princeton University. If an average of 80% of a neutron from a fission event is
captured by 238 U to make 239 Pu, the amount of Pu produced is
(411,000 tonne U)(0.711% − 0.667%)(0.8 239 Pu/235 U) = 145 tonne of Pu, (5.15)
which is the amount reported by the US Energy Information Agency (EIA). After
reprocessing, 380,000 tons of residual uranium were used as feedstock to obtain
enriched 235 U. The amount of 235 U separated is the difference of the input stream
(0.667%) and the tails stream (0.3%), which is 0.367% 235 U. The amount of 90%
enriched weapons-grade uranium produced is
(380,000 tonne)(0.00377)/(0.9 HEU) = 1500 tonnes of weapons-grade U, (5.16)
which is consistent with EIA’s reported value of 1400 tonnes.
5.4.2 Monitoring Warheads

Presidents Clinton and Yeltsin agreed in Helsinki in 1997 to explore the feasibility
of declaring and monitoring warheads in order to develop a comprehensive arms
control regime that could take the stockpiles to lower levels with greater confi-
dence. The hope was to control tactical warheads that had not been part of arms
control agreements up until then. In addition it could constrain surplus warheads
that could be covertly uploaded on missiles that had been downloaded as part of
START II.
5.4.3 Megatons to Megawatts

As the cold war world subsides from 70,000 warheads to a world of 4000 to 20,000
warheads under SORT, there is concern about Russia’s ability to manage its 150
tonnes of plutonium and 1000 tonnes of HEU. As an encouragement to Russia,
the United States placed 12 tons of plutonium and HEU under IAEA safeguards
in hopes that Russia would follow suit. The major barrier to weapons production
is availability of weapons-grade fissile materials, rather than the design and fab-
rication of weapons. Saddam Hussein spent $10 billion in his bid to produce the
weapons materials before the 1991 Gulf War stopped his program. On the other
hand, South Africa succeeded by spending only $200 million.
To reduce dangerous proliferation, the United States agreed to pay $12 billion
over 20 years to purchase 500 tonnes of HEU, after it is blended to low-enriched reac-
tor fuel. This arrangement gave needed funds to Russia at a critical time, preventing
P1: GFZ
5.4. Nonproliferation in the Former USSR 123
its bankrupt nuclear facilities from collapsing. The mixing process conserves both
total uranium mass and total 235 U mass:
mP = mHEU + mF and mP f P = mHEU f HEU + mF f F , (5.17)
235
where f is again the U fractional content of the named subscripts. To simplify
the mathematics with three streams and not four, the mixing of HEU is assumed to
be with natural uranium and not the actual 1.5% enriched form. Combining these
equations gives the mass of the product P, 4.4% enriched reactor fuel, from mixing
the natural uranium feed F to 90% enriched HEU:
mP = mHEU ( f HEU − f P )/( f P − f F ) = (500 tonne)(0.90 − 0.044)/(0.044 − 0.0071)
= 11,600 tonne. (5.18)
(The actual value is 15,000 tonnes with 1.5%-enriched uranium.) A 1-GWe reactor
has a core of 100 tonnes, of which one-third is refueled every 1.5 years, giving an
annual fueling of
(100 tonne/3)/(3/2 year) = 22 tonne/year. (5.19)
The product will fuel a 1-GWe reactor for a span of
(11,600 tonne)/(22 tonne/year) = 525 years, (5.20)
in agreement with the DOE estimate of 600 years. This amount will fuel the US’s
100 GWe of nuclear power for about 5 to 6 years. The amount of natural uranium
needed to denature HEU is
mF = mP − mHEU = (11,600 − 500)(tonne) = 11,000 tonne natural uranium.(5.21)
The cost of 4.4% LEU was determined above at $1000/kg. The approximate value
of the HEU deal is about
CRussian-LEU = (1.2 × 107 kg)($1000/kg) = $12 billion. (5.22)
However, the real situation is clouded because the United States is paying less
by only purchasing SWUs, and not the natural uranium feed. Alternatively, the
value/kilogram of HEU is
CRussian-HEU = $12 billion/500 tonne = $24,000/kg. (5.23)
The United States paid Kazakhstan $10 million for 600 kg of lightly protected HEU
at a cost of
CKazakhstan-HEU = $10 M/600 kg = $17,000/kg. (5.24)
5.4.4 Pu Buyout?
Uranium disposition is relatively easy since HEU is easily mixed to become LEU,
which has market value. On the other hand, plutonium cannot be denatured since
all Pu isotopes are fissile materials (Table 5.5). In addition, plutonium has a negative
monetary value because it is very costly to make MOX reactor fuel from it. A logical
P1: GFZ
Table 5.5. Isotopic composition of five grades of plutonium

[J.C. Mark, Science and Global Security 4, 111–128, 1993]
238 Pu 239 Pu 240 Pu 241 Pu 242 Pu
Super-grade – 98% 2% – –
Weapons-grade 0.012% 93.8% 5.8% 0.35% 0.022%
Reactor-grade 1.3% 60.3% 24.3% 9.1% 5.0%
MOX-grade 1.9% 40.4% 32.1% 17.8% 7.8%
Breeder blanket – 96% 4% – –
way for the United States to encourage Russia to dispose of its Pu would be for
Japan to buy Pu from Russia and not reprocess its own spent fuel. Japan continues
construction of a large reprocessing plant, even though it makes no economic sense.
The high cost of fabricating and storing plutonium fuel and the low cost of uranium
fuel gives plutonium “a zero or even negative commercial value,” in light water
reactors (LWRs) according to DOE (1992). In 2000, the United States and Russia
agreed to dispose of 34 tonnes of plutonium each, most likely by using it as MOX
fuel in reactors. The idea of immobilization of Pu in ceramics for underground
burial was abandoned by the second Bush administration.
5.5 Plutonium Production

5.5.1 Reactor-Grade and Weapons-Grade Pu
A typical LWR creates 0.5 239 Pu nuclei for every fission event. Some of the 239 Pu
nuclei are fissioned in the reactor, while others capture a neutron and become 240 Pu.
WgPu contains less than 6% 240 Pu after the fuel rods have remained in a reactor for
a few months (Section 1.5). However, fuel rods that remain in a reactor for a few
years produce RgPu containing over 20% 240 Pu. Spontaneous neutrons from 240 Pu
can preinitiate a nuclear explosion, greatly reducing its yield, but sophisticated designs,
unavailable to first-time programs, can overcome the preinitiation problem. Thus far, the
eight nations with plutonium weapons have chosen WgPu and rejected RgPu;
nevertheless, poor weapons of kiloton-size or good sophisticated weapons can be
made from RgPu. For this reason, the United States constrains reprocessing of US-
origin spent fuel and the IAEA maintains safeguards over both types of plutonium.
5.5.2 No Commercial Plutonium

During 1976–77, Presidents Ford and Carter called for an indefinite deferral of
spent-fuel commercial reprocessing, which was largely in response to India’s 1974
nuclear test with a Pu bomb. For decades, European nations and Japan have had
large plutonium programs for MOX fuels and breeders. The end of commercial
reprocessing precluded the use of plutonium in MOX fuels for thermal and breeder
reactors. The Ford/Carter decision was prompted by events in several smaller
P1: GFZ
5.5. Plutonium Production 125
countries, where governments tried to obtain reprocessing plants. These nations

used their small nuclear power programs to cover their motives, which was to make
nuclear weapons. The Carter policy of a once-through fuel cycle without reprocessing
gained adherents over time, as nations readjusted their views on the plutonium
economy, but reprocessing continues in UK, France, Japan, Russia, India and it will
soon begin in China. In the past, reprocessing plants had Pu losses of about 1%,
but they hope to reduce this to 0.1–0.2%. For a large 800-ton/year plant, a 1% loss
of Pu is (0.01 loss)(0.9% Pu)(8 × 105 kg spent fuel) = 70 kg/year and a 0.1% loss is
7 kg/year.
5.5.3 Energy in Spent Fuel

Let us compare the relative energy content of Pu and 235 U in 4.4%-enriched spent
fuel as compared to fresh fuel. A spent-fuel rod contains 0.8% 235 U (corrected for
build up of parasitic 236 U). If we assume 0.2% tails, the percent of 235 U in the spent
fuel available for reenrichment is about (0.8–0.2%)/4.4% = 13.6% of the fresh fuel
level.
Spent fuel has a content of 0.9% 239 Pu, which is less effective as a fuel than 235 U in
reactors because only 74% of captured thermal neutrons fission 239 Pu as compared
to 86% for 235 U (with the remainder producing 236 U and 240 Pu). These branching
ratios reduce the energy available from 239 Pu by 0.74/0.86 = 86% as compared to
an equivalent amount of 235 U. From this, the fractional energy value of plutonium
in spent fuel in terms of 235 U content in fresh fuel is
(0.9% Pu)(0.86 239 Pu/235 U energy)/(4.4% 235 U) = 17.6%. (5.25)
If the uranium in the spent fuel could be used, this would give an additional
13.6%. However this uranium contains considerable 236 U, which as a neutron par-
asite would not be useful in MOX, which uses natural uranium. The uranium in
the spent fuel could be used for feed for an enrichment plant, but it would be
bothersome as the uranium is quite radioactive. The poor quality uranium and the
MOX tasks of reprocessing and fuel fabrication are more expensive than making
new fuel from natural uranium.
The economics of using spent fuel for MOX is considerably worse for small
reprocessing plants. A large plant can annually accommodate 800 tons of spent
fuel, while a smaller nation might purchase a plant that can process 100 tonnes.
Since a 1-GWe nuclear plant annually produces about 22 tonnes of spent fuel,
the 800 tonne/year plant accommodates 35 GWe nuclear power, while the 100
tonne/year size accommodates only 4 GWe . The annual cost to purchase, operate,
and maintain a large facility scales with its size. Economic analysis typically uses
costs that scale with capacity to the 0.6 power. This raises the cost over time for a
small plant as compared to a large plant by factor of
(100 tonne/year/800 tonne/year)0.6 /(100 tonne/800 tonne) = 2.3. (5.26)

P1: GFZ
Thus, the reprocessing cost for a small plant is two to three times higher in smaller
countries. In Chapter 16, we examine the economic trade-off between breeder re-
actors and the once-through fuel cycle for LWRs.
5.5.4 Uranium Supplies

The United States has tried to dissuade other counties from entering the plutonium
economy of MOX and breeders by stretching uranium supplies. Let us examine
some ways in which this is being done. If 4.4%-enriched fuels are used, 22 tonnes
of spent fuel are annually removed from 1-GWe reactors. This amounts to a re-
quirement of 1 ton/year of 235 U fuel, from
(22 tonne/GWe -year)(4.4% 235 U) = 1 tonne 235 U/GWe -year (5.27)
at a thermodynamic efficiency of 1/3. Over a 40-year life, a reactor consumes

(22 ton/year)(40 year) = 900 tonnes of 4.4%-enriched uranium. In Section 5.3 it
was shown that it takes 8.2 kg of natural uranium to make 1 kg of 4.4%-enriched
uranium, a process that requires 7400 tonnes of natural uranium for a 40-year life
with the once-through fuel cycle (900 × 8.2). The United States has 3 million tons of
reasonably assured reserves and speculative resources, which could accommodate
(3 Mtonne)/(1 GWe -life/7400 tonne) = 400 GWe -lifetimes. (5.28)
In the 1970s, these reserves were considered minimal because US nuclear power
was expected to rise to a level of more than 1000 GWe by the year 2000. At the end
of the century, the US nuclear power peaked at 100 GWe , which will consume 25%
of the 3-Mton reserves. There is additional uranium available in low-grade ores,
as well as in “mining” the enrichment tails to 0.05%, and in dismantled nuclear
weapons and in ocean water. The shift from the once-through cycle to the breeder
has stopped in an era of uranium surplus, curtailed nuclear orders, and inexpen-
sive combined-cycle gas turbines powered with natural gas. (See Section 16.8 for
plutonium economics.) Nuclear power might have a role in the future because of
potential climate change problems or as natural gas supplies become limited, but
this will only happen if the US government enters the marketplace.
5.5.5 Longer Burn-Up Fuels

Uranium requirements can be reduced by 10% if fuel remains in reactors for a
longer residency. The longer time allows a greater fraction of 235 U nuclei to fission
and it allows more in-situ 239 Pu production and fission. A measure of fuel residence
time is the “burn-up,” which is thermal energy produced per unit mass in units of
MWt -day/tonne where Wt is thermal power. A 3.2% fueled 1-GWe reactor, with a
thermal power of 3 GWt and core of 100 tonnes has a fuel burn-up of
(3000 MWt )(3 year)(365 day/year)(0.8 load)/(100 tonne) = 30,000 MWt -day/tonne.
(5.29)
P1: GFZ
5.5. Plutonium Production 127
Improved fuels with an increased burn-up of 45,000 MWt -day/tonne remain

in reactors for 4.5 years, reducing spent fuel storage by 30%, compared to 3.2%
enriched (30,000 MWt -day/tonne, 3 years). A 10% savings in uranium with 4.4%
fuel takes place because less 235 U is left in tails and spent fuel. About 0.8% of spent
fuel heavy metal is 235 U. The annual waste of 235 U per GWe -year with 3.2% enriched
is
(100 tonne/3 year)(0.8%) = 0.27 tonne-235 U/year, (5.30)
as compared to 4.4% enriched fuel,
(100 tonne/4.5 year)(0.8%) = 0.18 tonne-235 U/year. (5.31)
A saving of 0.1 tonne 235 U/year stretches uranium supplies by 10% since a large
reactor consumes 1 tonne 235 U/year.
5.5.6 Research Reactors

Research reactors provide two paths to nuclear weapons. The first path is the di-
version of HEU reactor fuel. The United States has worked to close this path by
reducing 235 U content in the research reactor fuel from 90% to 20%. The same re-
actor power can be maintained in most cases by increasing uranium density while
reducing the 235 U-enrichment level, which has the effect of maintaining the same
volume density of 235 U.
At the other extreme, Israel’s Dimona and India’s Cirus research reactors use
natural uranium fuel and a heavy water moderator to produce plutonium at the
rate of 0.3 kg/MWt -year. The time needed to obtain 5 kg for a Pu warhead from
the 40-MWt reactor is
t = 5 kg(1 MWt -year/0.3 kg)(1/40 MWt ) = 0.4 year, (5.32)
a rate that allows production of enough plutonium for two warheads a year. The
United States urged conversion of natural uranium research reactors to a higher
enrichment level of 20% to substantially reduce plutonium production. The rate of
plutonium production depends on the ratio of fertile 238 U to fissile 235 U. The ratio
of fertile/fissile in natural uranium fuel (0.7% 235 U) fuel is
238
U/235 U = 99.3%/0.7% = 140, (5.33)
while the fertile/fissile ratio in 20% fuel is

238
U/235 U = 80%/20% = 4. (5.34)
By switching from natural U fuel to 20% fuel, the Pu production rate is reduced
from two warheads per year by a factor of 35 to one warhead in 15 years, greatly
increasing the time to make a warhead. Since Israel and India had their first war-
heads appear in 1968 and 1974, respectively, they could each have produced some
75 warheads by 2000 using natural uranium.
P1: GFZ
5.5.7 CIVEX
Walter Marshall and Chauncey Starr proposed in 1978 a fuel cycle that would pro-
vide a radiation barrier to prevent terrorists from working with stolen Pu. A high
concentration of 238 Pu in plutonium increases the radiation rate, increases ther-
mal power, and increases the rate of spontaneous neutron emission. The practical
complications of a CIVEX cycle precluded its adoption, but it raised interesting
questions. The use of the 88-year half-life of 238 Pu is long enough to allow Pu to be
protected, but short enough to produce considerable radioactivity and heat. If 1%
238
Pu is added to 5 kg of Pu, there will be a heating rate of
(5 kg Pu)(1% 238 Pu)(560 W/kg 238 Pu) = 28 W. (5.35)
If the plutonium is weapons-grade Pu, there will be an additional 5 kg × 2.3 W/kg
= 12 W from 240 Pu for a total of 40 W. This heat source raises the temperature of a
bare sphere (Chapter 11 on heat transfer) by
P = (1/R)AT = (1/R)(4πr 2 )(T), (5.36)
where P is power lost through a surface area 4πr 2 , R is thermal resistivity, and
T is temperature difference. Using a convection, radiation and high-explosive
insulation R-value of 0.4 W/m2 -◦ C, the temperature rise at the surface of a 10-cm
radius sphere made with CIVEX is
T = (RP)/(4πr 2 ) = (0.4 W/m2 -◦ C)(40 W)/(4π )(0.1 m)2 = 130◦ C. (5.37)
This temperature is too high for maintaining explosives over long time periods. Such
an outcome could be mitigated with thermal bridges and fins, but the device would
be more complex.
A less sophisticated weapon made without CIVEX, but with 8 kg of reactor grade
Pu would produce 8 kg × 10.5 W/kg = 90 W. The temperature rise might be
T = (RP)/(4πr 2 ) = (0.4)(90 W)/(4π )(0.1)2 = 200◦ C, (5.38)
which is similar to estimates by J. Carson Mark.
5.6 MTCR and Scuds

The first theater ballistic missile, the German V2, had a payload of 900 kg and a
range of 300 km. Fifty years later, Scud-type missiles, similar to the V2, had pro-
liferated to 25 nations, which were the type launched by Iraq in the 1991 Gulf
War. The issue of missile proliferation encouraged establishment of the Missile
Technology Control Regime (MTCR) in 1987. MTCR constrains exports of mis-
siles and their subsystems that are capable of delivering more than 500 kg, the
size of a crude nuclear warhead, at a Scud range of 300 km. In 1993 the throw-
weight criteria was removed since biological weapons are much lighter and can
be as dangerous. There are exemptions in MTCR for transfer of space launch
P1: GFZ
5.6. MTCR and Scuds 129
Figure 5.5. Payload and range of aircraft and missiles. Aircraft can carry larger payloads
than Scud missiles at theater ranges of 1000 km. As indicated in the graph, aircraft can be
more lethal, but they can be more vulnerable to attack, and they respond less quickly (Office
of Technology Assessment, 1993).
vehicles (SLV) for nonmilitary uses. These exemptions complicate MTCR compli-
ance. Since MTCR is a quasi-executive agreement and not a signed treaty, it has lead
to many misunderstandings, such as the one that lead to the Chinese export of M-11
missiles to Pakistan. Payload and range of aircraft and missiles are displayed in
Fig. 5.5.
Throw-weight and range are linked, since a reduction in throw-weight mass m
allows an increase in velocity v and range R. If fly-out energy is constant (ignoring
the rocket equation), the ratio of velocities and masses is
(v2 /v1 )2 = m1 /m2 . (5.39)
The ratio of ranges for a flat Earth is proportional to the energy, giving
R2 /R1 = (v2 /v1 )2 = m1 /m2 . (5.40)
From this, a 1% reduction in mass increases the range by 1% without the rocket
equation. On the other hand, if the impulse to the payload is constant, that
is v1 m1 = v2 m2 , a 1% reduction in mass increases velocity by 1% and range
by 2%.
P1: GFZ
5.6.1 Scuds
In the MTCR Annex there are additional export constraints. MTCR exports of rocket
engines must not exceed a total impulse capacity of 2.5 × 105 pound-seconds (note,
SI units are not used in an international agreement). To see if this figure is consistent
with a 500-kg mass traveling at range of 300 km, consider the following: The flat
Earth range is
R = 300 km = 3 × 105 m = v2 /g = v2 /10 m/s2 , (5.41)
for a fly-out velocity of
v = (3 × 106 )1/2 = 1730 m/s = 1.7 km/s. (5.42)
Using the reported Scud velocity of 2 km/s, the impulse imparted to a 500-kg
projectile is
mv = (500 kg)(2000 m/s) = (106 N · s)(1 lb/4.5 N) = 2.2 × 105 lb · s, (5.43)
close to the 2.5 × 105 lb s limit. The impulse must be increased to take into account
the momentum lost to the missile body and atmospheric drag. As is the case with
all flat Earth trajectories, flight time is twice the time it takes the missile to reach
the top of its trajectory:
t = 2vo sin 45◦ /g = 2(2000 m/s)(0.707)/(10 m/s2 ) = 280 s = 4.7 min, (5.44)
which agrees with the published value of 5 min. The height of a Scud trajectory is
h = g(t/2)2 /2 = (10 m/s2 )(280 s/2)2 /2 = 100 km. (5.45)
Problems
5.1 The Baruch Plan. What were the Baruch plan proposals, as read in its text?
Why was the plan not acceptable to the Soviets? What were its advantages
and disadvantages?
5.2 Eisenhower’s Atoms for Peace speech. What were the advantages and disad-
vantages of the actions that resulted from President Dwight D. Eisenhower’s
speech and policies? How is this affected by advances in technology? In what
ways did it promote civil nuclear power and nonproliferation?
5.3 Nuclear Nonproliferation Treaty. What are the trade-offs and responsibilities
for the NWS and the NNWS in NPT Articles 1–6? Under what conditions do
you think the NWS are committed to be assured suppliers of nuclear fuel?
5.4 Soviet Pu production. The Soviets placed 410,000 tons of natural uranium
in reactors to produce plutonium. (a) How much Pu is produced if natural
uranium’s 235 U content of 0.711% is reduced in the reactor to 0.677%, and 0.8
239
Pu nuclei are created for every 235 U that fissions? (b) Why is Pu found on
the ground near US enrichment plants?
P1: GFZ
Problems 131
5.5 Soviet HEU production. The U from Pu production reactors was reprocessed
and used as feedstock for enrichment. How much 90% enriched HEU was
produced from 380,000 tons of U obtained from 0.677% 235 U with 0.3% tails?
5.6 1 tonne 235 U = 1 GWe -year. Starting with energy of 200 MeV per fission,
show that a power plant producing 1 GWe at 33% thermal efficiency consumes
1 tonne 235 U/year. (Some power is gained from Pu made in the reactor from
238
U, but this is offset by the 235 U and 236 U remaining in spent fuel.)
5.7 U/GWe -life. How much natural uranium does a 1-GWe plant consume using
3.2% fuel over a 30-year life? Assume efficiency of 1/3 and 0.2% tails.
5.8 Mining tails. Russia is using quality centrifuges to lower tails to 0.05%. If
Russia mines 350,000 tons of 0.3% uranium tails, how many tons of 4.4% fuel
can be produced? How many GWe -years of fuel can be obtained?
5.9 Mass difference and mass ratio. What is the one-stage separation factor for
separating 6 Li from 7 Li with (a) thermal-diffusion, (b) gaseous-diffusion, and
(c) centrifuges, using text parameters?
5.10 Li laser isotope separation. (a) Show that the isotope shift frequency is pro-
portional to the difference of the root mean square of the nuclear radii of
235
U and 238 U. (b) What is the isotope shift between 6 Li and 7 Li? Use hydro-
genic wave functions and a nuclear radius of 1.4 A1/3 fermi (see Hafemeister,
1980).
5.11 235 U from LIS. What is the electrostatic energy difference (and frequency shift)
for outer s-electrons around 235 U and 238 U? Use a nuclear radius of 1.4 fm
A1/3 fermi and an electron density at the nucleus of an outer electron of 5 ×
1026 /cm3 .
5.12 Value function. What is the thermodynamic value, V( f ) = (2 f – 1) ln( f /
1 – f ), for a gas containing 0%, 25%, 50%, 75%, and 100% 235 U with the re-
mainder 238 U?
5.13 Reactor fuel and HEU for weapons. (a) Show that it takes 11 kg of natural
uranium and 7 kg to make 1 kg of 4.4% fuel, with 0.2% tails. (b) What does it
cost to make 1 kg if SWUs cost $100 and natural uranium costs $30/kg. (c) How
many SWUs, in kilograms of natural uranium and dollars does it take to make
1 kg of 93.3% enriched? (d) How many SWUs are needed to operate a 1-GWe
plant for a year? How many weapons could be made with this SWU amount?
(e) What fraction of separative work is saved by making 90% HEU from 4.4%
enriched starting material?
5.14 Research reactor fuel. (a) By what factor should uranium density be increased
to maintain power if enrichment is dropped form 90% to 20%? (b) What is the
fractional savings in SWU, in kilograms of natural uranium and in dollars, by
shifting from 90% to 20% at 0.2% tails?
5.15 Breeder capital cost versus uranium price. Breeder reactor capital costs are
perhaps 50% higher than capital costs for an LWR. However, LWRs use more
natural uranium and enrichment services than do breeders. If an LWR uses
22 tonnes per year of 4.4% fuel, what is the cost per year if natural uranium
costs $33/kg and a kg-SWU costs $100? On a breakeven basis, how much can
the breeder capital cost exceed the capital cost of the LWR if the carrying cost is
P1: GFZ
10%/year? (This ignores reprocessing and MOX fabrication costs. See Section
16.8.)
5.16 Monitoring spent fuel rods. A country removed 250 of its 1000 fuel assemblies
that are under IAEA safeguards. (a) What is the probability that an inspection
of one assembly discovers a violation? (b) Show that the probability of a vi-
olation being discovered with n inspections is P = 1 – (1 – f)n with fraction
of violations f = V/N, where V is number of violations and N is number of
declared fuel assemblies. (c) What is the probability of detection for 1, 3, and
5 inspections?
5.17 Mton to MW. The text assumed Russia mixed natural uranium with the
500 tons of 90% HEU, but they use 1.5% enriched to avoid health problems
from excess 234 U. (a) How many tons of 4.4% fuel result from a 1.5% feed?
(b) How many GWe -year fuel will this provide?
5.18 USEC problems for HEU-Russia. The privatization of the US Enrichment
Corporation complicates funds for Russia since USEC wants to return an
equivalent amount of natural uranium. Using the values given in the text,
what fraction of the $12 billion is for uranium, and what fraction is for the
enrichment services?
5.19 Super burn-up fuel. Scientists hope to create robust fuels that can sustain a
burn-up of 60,000 MWt -day/tonne. (a) How much natural uranium would
be saved compared to 30,000 and 45,000 MWt -day/tonne fuels for the US’s
100 GWe ? (b) What is the fractional reduction in spent fuel?
5.20 World Trade Center (9-11-2001). Building structures are constructed to carry
twice the weight that is above each floor. If flames weaken the columns until
they collapse, dropping the material above 3.5 meters, what force is needed
to stop the falling material in 5 cm? How does that compare to the maximum
support force? What is the deceleration?
5.21 October 2001 anthrax. Five people died and 22 were sickened with anthrax-
laden letters. The United States purchased several 10-MeV, 18-kW electron
accelerators that can sanitize mail at 570 kg/h. Does this dose rate seem rea-
sonable?
Bibliography
Albright, D. and H. Feiveson (1988). Plutonium recycling and the problem of nuclear pro-
liferation, Ann. Rev. Energy Environ. 13, 239–266.
Albright, D., F. Berkhout and W. Walker (1997). Plutonium and Highly-Enriched Uranium: 1996
World Inventories, Capabilities and Policies, Oxford Univ. Press, Oxford.
Bodansky, D. (2004). Nuclear Energy, American Institute of Physics Press, New York.
Bukharin, O. (1996). Analysis of the size and quality of uranium inventories in Russia, Sci.
Global Secur. 6, 59–77.
Bukharin, O. (1996) Security of fissile materials in Russia, Ann. Rev. Energy Environ. 21,
467–498.
Bunn, M. and J. Holdren (1997). Managing military uranium and plutonium in the United
States and the former Soviet Union, Ann. Rev. Energy Environ. 22, 403–486.
P1: GFZ
Bibliography 133
Cirincione, J. (2002). Deadly Arsenals: Tracking Weapons of Mass Destruction, Carnegie Endow.
Int. Peace, Washington, DC.
Craig, P. and J. Jungerman (1990). Nuclear Arms Race Technology and Society, McGraw Hill,
New York.
Hafemeister, D. (1980). Science and Society Test V: Nuclear Nonproliferation. Am. J. Phys.
48, 112–120.
Krass, J., et al. (1983). Uranium Enrichment and Nuclear Weapons Proliferation, Taylor and
Francis, London.
Lamarsh, J. (1977). Introduction to Nuclear Engineering, Addison Wesley, Reading, MA.
Mark, J.C. (1993). Explosive properties of reactor-grade plutonium, Sci. Global Secur. 4, 111–
128.
National Acad. of Sciences (1994). Management and Disposition of Excess Weapons Plutonium,
National Academy Press, Washington, DC.
National Research Council (2002). Making the Nation Safer: The Role of Science and Technology
in Countering Terrorism, National Academy Press, Washington, DC.
Nero, A. (1979). A Guidebook of Nuclear Reactors, University of California Press, Berkeley, CA.
Office of Technology Assessment (1977). Nuclear Proliferation and Safeguards, OTA, Washing-
ton, DC.
———(1993). Proliferation of Weapons of Mass Destruction, OTA, Washington, DC.
———(1993). Technologies Underlying Weapons of Mass Destruction, OTA, Washington, DC.
———(1993). Dismantling the Bomb and Managing the Nuclear Materials, OTA, Washington,
DC.
———(1995). Nuclear Safeguards and the IAEA, OTA, Washington, DC.
Scheinman, L. (1987). The International Atomic Energy Agency, Resources for the Future, Wash-
ington, DC.
P1: GFZ
Environment
6 Air and Water Pollution ................................................................... 137

6.1 Acid Rain pH............................................................................. 138
6.2 Clean Air Act and Allowance Trading............................................ 138
6.3 Pollution Scaling ........................................................................ 141
6.4 Power Plant Plumes .................................................................... 144
6.5 Automobile Emissions in the LA Air Basin ..................................... 148
6.6 Stratospheric Ozone.................................................................... 150
6.7 Purifying Water.......................................................................... 153
6.8 Environmental Chemistry ............................................................ 156
6.9 Flowing Water............................................................................ 158
7 Nuclear Pollution ............................................................................ 163

7.1 Low-Dose Radiation ................................................................... 163
7.2 Loss-of-Coolant Reactor Accidents ................................................ 169
7.3 Plume of 137 Cs from a LOCA ........................................................ 173
7.4 Weapon Accident Plutonium Plume .............................................. 177
7.5 Dirty Bombs .............................................................................. 180
7.6 Fault Tree Analysis ..................................................................... 181
7.7 Geological Repositories ............................................................... 185
7.8 Indoor Radon............................................................................. 190
8 Climate Change .............................................................................. 197

8.1 Introduction .............................................................................. 197
8.2 CO2 Projections .......................................................................... 201
8.3 Upper-Atmospheric and Surface Temperatures ............................... 205
8.4 Temperature Refinements ............................................................ 210
8.5 Link Between CO2 and Temperature.............................................. 213
8.6 Solar and Oceanic Variations ........................................................ 216
135
P1: GFZ
8.7 Heat Islands............................................................................... 221

8.8 Policy Options ........................................................................... 224
9 Electromagnetic Fields and Epidemiology .......................................... 233

9.1 Power Line Health Effects?........................................................... 233
9.2 Epidemiology ............................................................................ 238
136
P1: GFZ
6
Air and Water Pollution
Many pollutants enter our bodies through air we breath at 13,000 l/day and water
we drink at 2 l/day. Our goal is to quantitatively understand the basic environmen-
tal physics of pollution, leaving more detailed discussion to others. This chapter
covers the following topics: acid rain, power plant plumes, pollution allotment
trading, pollution scaling laws, time-dependence of air pollution, auto emissions
in the LA air basin, ozone depletion, water purification, environmental chemistry,
and time dependence of pollution in lakes.
Because of the 1970 Clean Air Act, air quality improved by 29% from 1970 to
2000 for the six main air pollutants (CO, NO2 , O3 , SO2 , Pb, particulates). This
reduction occurred over a period when population increased 36%, vehicle-miles
traveled increased 143% and Gross Domestic Product increased 158%. Yet, in spite
of the pollution reductions, Pasadena residents often cannot see the San Gabriel
Mountains, and many cities exceed atmospheric ozone standards. The Environ-
mental Protection Agency concluded that Americans living in heavily populated
areas have a 12% greater risk of dying from lung cancer than residents living in
less populous regions. Each year 30,000 Americans (600,000 globally) die from dis-
eases related to air pollution, which is a 16% reduction from 1982 data. Americans
spend 86% of their time indoors, 6% in vehicles, and 8% outdoors. According to
EPA estimates there are 3000 lung cancer deaths per year from secondhand smoke
and 14,000 from radon.
Scarcity of water is a rarely discussed aspect of the Israeli Palestinian conflict
for control of the West Bank. In the United States, water overuse lowered Ogallala
aquifer in the Plains states by 3.6 million acre feet/year during the 1990s. (An acre-
foot is the amount of water that covers one acre a foot deep.) This issue is not
confined to the Southwest and the Plains, as Maryland aquifers dropped by 40 feet
in 20 years. Such loses could be reduced if wasteful practices were abandoned.
California’s alfalfa crop, for example, gets 25% of its irrigation water from flooding
Imperial Valley desert fields at cheap prices. Pollution enters drinking water and
estuaries through agriculture runoffs. Four kilogram of nitrogen in fertilizers is
needed to produce 10 kg of food, but 90% of the nitrogen is wasted in runoff.
Nitrogen fertilization must be better managed if estuaries are to remain vibrant
fisheries.
137
P1: GFZ
138 6. Air and Water Pollution
6.1 Acid Rain pH

Acid rain attacks forests and kills fish. It is primarily the result of sulfur and nitrogen
emissions from coal-burning factories and power plants and from cars. In some
cases, acidity is as strong as vinegar (pH = 2.4). This can be compared to rain from
pristine air, which contains CO2 at 360 ppm, giving a 5.5 pH, as CO2 becomes
H2 CO2 . As they seep into the ground, both sulfur and nitrogen acids leach calcium
and other trace minerals from the ground. The harm done by acid rain can be
cumulative, as lakes and soils that absorb acid rain lose their ability to buffer acids.
The good news is that US industries cut sulfur emissions by more than 50% since
1980 by using lower sulfur coals. Further reductions come from using scrubbers
that combine sulfur dioxide (SO2 ) with water vapor containing crushed limestone
(calcium carbonate, CaCO3 ):
SO2 + CaCO3 ⇒ CaSO3 + CO2 . (6.1)
Let us estimate the pH of acid rain in 1980 from US coal consumption of 600 mil-
lion tons of coal containing an average of 1% sulfur. We assume that most acid rain
falls on the eastern United States with an area of 0.5 million square miles, giving
an annual sulfur deposition of
(1% S)(600 Mton)/(0.5 × 106 mi2 ) = 12 ton S/mi2 = 4.5 g S/m2 . (6.2)
The number of H2 SO4 moles deposited per square meter is
(4.5 g S/m2 )/(1 mole H2 SO4 /32 g S) = 0.14 H2 SO4 mole/m2 . (6.3)
Each H2 SO4 molecule contributes two hydrogen ions and NOx compounds, in
effect, add an average of one more hydrogen ion to give a density of 0.42 H−
mole/m2 . Average northeast rainfall is 0.9 m per year giving an annual rainfall
density of (l = liter)
(0.9 m)(103 l/m3 ) = 900 l/m2 . (6.4)
From this, the pH of eastern rain is
pH = − log[H− moles/l] = − log[0.42 moles/900 l] = 3.3, (6.5)
which is much smaller than the EPA 1980 average pH of 4.4.
Our calculation indicates a higher acidity since it does not take into account
that tall smokestacks disperse acid rain to Canada and the Atlantic Ocean. (Maine
Governor Joseph Brennan complained that Maine was “at the end of the geographic
exhaust pipe.”) Acidity is further reduced as some sulfur settles out as small parti-
cles and aerosols. The sulfur problem has improved over time, but NOx problems
have worsened.
6.2 Clean Air Act and Allowance Trading

Air quality standards enforced by the EPA are given in Table 6.1. Included is the lat-
est controlled pollutant, particulate matter under 2.5 μ in diameter with an annual
average limit of 15 μg/m3 .
P1: GFZ
6.2. Clean Air Act and Allowance Trading 139
Table 6.1. National Ambient Air Quality Standards for 2002.

Averaging
Pollutant time (hours) Concentration Peak LA
Carbon monoxide (CO) 1 35 ppm (40 mg/m3 )
8 9 ppm (10 mg/m3 ) 10 ppm
Nitrogen dioxide (NO2 ) annual mean 0.053 ppm (0.1 mg/m3 ) 0.044 ppm
24 0.65 ppm (1.2 mg/m3 )
Ozone (O3 ) 1 0.12 ppm (0.24 mg/m3 ) 0.17 ppm
8 0.08 ppm (0.16 mg/m3 ) 0.11 ppm
Sulfur dioxide (SO2 ) annual mean 0.03 ppm (0.08 mg/m3 ) 0.003 ppm
3 0.5 ppm (1.3 mg/m3 )
24 0.14 ppm (0.365 mg/m3 ) 0.010 ppm
Particulates < 10 μ diameter annual mean 50 μg/m3 46 μg/m3
24 150 μg/m3 93 μg/m3
Particulates < 2.5 μ diameter annual mean 15 μg/m 24 μg/m3
24 65 μg/m3 83 μg/m3
Lead (Pb) calendar quarter 1.5 μg/m3 0.06 μg/m3
Criteria for setting EPA standards are based on an Air Quality Index of 100, which is the
lower limit for “unhealthy for sensitive groups.” The last column gives the peak air quality
data for Los Angeles in 2000. Bold type indicates violations for carbon monoxide, ozone and
particulate matter (< 2.5 μ diameter).
6.2.1 SO2 ppm Levels

Hospital admissions rise when SO2 exceeds 0.1 ppm (by volume) for 4 days or
1.0 ppm for 5 min. An extreme example of this was found in London’s 1952 inver-
sion, when the air held an SO2 level of 0.7 ppm for 3 days, causing at least 3000
excess deaths, mostly among the elderly. The public health Air Quality Standard
is 0.14 ppm over 24 h. (See Fig. 6.1 for data on the London killer fog.)
6.2.2 Allowance Trading

The linchpin of EPA’s mitigation acid rain is allowance trading, which allows one in-
dustry to buy and sell SO2 emission allowances with other industries, or to buy and
sell allowances on the open market. Allowance trading characterizes sulfur emis-
sion allowances as marketable commodities between utilities. If Utility A finds its
sulfur mitigation options are too expensive, it can pay another industry to mitigate
its emissions while continuing its own operations unchanged, thus staying in com-
pliance with the Clean Air Act. Utilities decide on the basis of cost effectiveness on
how to proceed to lower their SO2 emissions through a combination of allowance
trading and the following measures: Energy conservation; increased reliance on
renewable energy; reduced production, installation of pollution controls; switch-
ing to lower sulfur fuels; or developing other strategies. (Similar carbon allowance
trading is part of the Kyoto–Bonn climate modification protocol.)
The unit for sulfur allowance trading is pounds of SO2 per million Btu. Changes
in technology lowered SO2 emissions from 4 lb/MBtu of fuel in 1980 to less than
2 lb/MBtu. A coal plant typically burns 10,000 Btu to produce a kWh of electrical
P1: GFZ
Figure 6.1. London killer fog. Daily mean pollution concentration and daily number of
deaths during December 1952. An inversion layer trapped London’s dirty coal atmosphere
and killed over 3000 people (Wilson and Crouch, 2001).
energy. The 4-pound rate gives a sulfur rate of

(4 lb SO2 /MBtu)(10,000 Btu/kWh) = 4 lb SO2 /100 kWh. (6.6)
Since it takes a little less than a pound of coal to make 1 kWh, the above rate
converts to about 4 lb SO2 /100 lb coal. The 1980 norm was basically determined
from the emissions of 2% sulfur Eastern coal. Since oxygen is 50% of the weight
of SO2 , 100 lb of 2%-sulfur coal make 4 lb of SO2 for 100 lb of coal, which was the
1980 unit. EPA now uses a lowered allowance rate of 2.5 lb SO2 /MBtu, since coals
with less sulfur are now being used. As an example, a 100-W bulb over a 10-h night
uses 1 kWh/night and 365 kWh/year. At the 4 lb/MBtu standard, this added 15 lb
SO2 /year to the atmosphere, while the new standard lowers this to 9 lb SO2 /year.
EPA tried to force 32 midwest coal plants to install modern scrubbers, a bid that
became a political issue. The 1970 Clean Air Act grandfathered old coal-fired plants,
exempting them from installing antipollution equipment, which newer plants were
required to have. Yet grandfathered plants have been modernized since 1970. The
legal case turned on (1) the degree of modernization of the plants toward the status of a
“new” plant and (2) how much status to give the grandfather clause. The midwestern
P1: GFZ
6.3. Pollution Scaling 141
states challenged EPA by calling for continued grandfathering of older plants,

while the eastern states argued the converse. The March 2000 the US Court of
Appeals favored the east, ruling that midwest plants must be modified or shut
down. However, in August 2003 President Bush allowed exemptions from the
requirement to reduce emissions for plants that modernized less than 25% of the
total capital cost of a new power plant.
6.3 Pollution Scaling

City size can be used to calculate relative levels of pollution in cities. Scaling calcu-
lations show that large cities have higher concentrations of polluted air than small
cities. (Scaling laws in Section 1.3 gave estimates for critical masses of nuclear
weapons.)
Consider a small city Alpha with an SO2 level of 0.015 ppm over its area of
5 km × 5 km. For simplicity, we assume SO2 is emitted from many small sources
at a constant source rate per unit area s, measured in kg/km2 . Through scaling laws
we will determine the concentration c of SO2 in a larger city Beta, whose area is
50 km × 50 km.
The scaling factor between two cities is the ratio of their lengths L. The scaling
factor of Beta with respect to Alpha is L β /L α = 50 km/5 km = 10 = n. Beta’s total
pollution source rate Sβ in kg/s is n2 times as great as Alpha’s Sα , or
Sβ = L β 2 s = n2 L α 2 s = n2 Sα . (6.7)
Pollutants cool as they rise until they reach the inversion height H, the level
above the ground where pollution temperature equals atmospheric temperature.
We assume cities Alpha and Beta have the same inversion height. Wind blows
pollution away from Alpha through a cross-sectional area L α H, while Beta has a
cross-sectional area that is n times larger, L β H = nLα H. (See Fig. 6.2.) Steady state
Figure 6.2. Scaling laws for urban pollu-

tion. The wind blows contaminated air with
velocity u through an area of LH, the lat- L
eral width of the city L times the inversion
height H.
P1: GFZ
conditions exist when pollution production Sproduction equals wind-driven removal,

Sremoval . The flux f of pollution at the trailing edge of a city is
f = uc (6.8)
2
in kg/s m , where u is wind velocity and c is SO2 pollution concentration.
The total rate of pollution removal from a city is the exiting flux times inversion
area,
Sremoval = f × area = (uc)(L H). (6.9)
Equating the production and removal rates for Alpha and Beta gives
Sα-prod = Sα-removal = uc α L α H (6.10)
Sβ-prod = Sβ-removal = uc β L β H = uc β nL α H. (6.11)
Since Sβ-production = n2 Sα-production , it follows that Beta’s pollution is 10 times greater
than Alpha’s at
c β = nc α = 10 × 0.015 ppm = 0.15 ppm. (6.12)
Beta’s larger size raises its SO2 level over the Primary Air Quality Standard of
0.14 ppm. Therefore factories and utilities in Beta must mitigate SO2 with scrubbers,
fuel-switching or fuel conservation, while Alpha may continue its usual practices.
Fig. 6.3 shows this scaling effect on lead blood levels in cities between 0.1 and
10 million.
6.3.1 Time-Dependent Air Pollution

Pollution levels rise and fall during the day in sync with periodic pollution emis-
sions and solar photochemical reactions. Scaling laws reveal that large cities are
slower to reach equilibrium than cities of smaller size. The rate at which a city adds
pollution to the atmosphere,
Sprod = sprod L 2 , (6.13)
when the pollution production rate of the city Sprod is not balanced with the removal
rate,
Sremoval = −ucLH, (6.14)
because pollution stored in the air is always changing. The flow of pollution into
the air is the product of air volume times rate of change of SO2 concentration,
Sair = (air volume)(dc/dt) = L 2 H(dc/dt). (6.15)
Combining the three rates of change gives
Sair = Sprod + Sremoval (6.16)
L H(dc/dt) = sL − ucLH.
2 2
(6.17)
Rearranging gives
dc/dt = s/H − uc/L . (6.18)
P1: GFZ
6.3. Pollution Scaling 143
Figure 6.3. Blood lead concentration versus population. Blood lead levels are compared to
city sizes. The lead comes from the sale of leaded gasoline (Thomas, et al., 1999).
This differential equation is identical to that of a battery charging a capacitor

through a resistor. Integrating the equation gives pollution concentration as a func-
tion of time t:
c(t) = sL/uH[1 − e −ut/L ]. (6.19)
The time for pollution to traverse the full length of the city is the replacement
time, τ = L/u, which is also the relaxation time τ . A 5-m/s wind blowing over a
P1: GFZ
small town of 5 km × 5 km has a fast relaxation time of

τ = L/u = (5000 m)/(5 m/s) = 1000 s = 0.3 h. (6.20)
A large city of 50 km × 50 km has τ = 3 h. If the wind drops to 1 m/s, τ is 30 h
with considerable pollution left over to contribute to the following day. Rewriting
c(t) using τ gives
c(t) = sL/uH[1 − e −t/τ ] = c max [1 − e −t/τ ]. (6.21)
Initially c(0) = 0, when pollution began. For very long times c(∞) = c max , the
maximum concentration level. Concentration reaches its asymptotic level c max =
sL/uH at steady state when production and removal are exactly equal and atmo-
spheric concentration is constant (dc/dt = 0). If the wind drops from 2 m/s to
0.2 m/s and the inversion level drops from 1000 m to 100 m, maximum concentra-
tion rises by a factor of 100! If a city has 10 times the source production rate s and
it is 10 times larger along one side than another city, this results in another anther
factor of 100. Of course, local geography is an important factor. Pollution in the Los
Angles air basin is complicated by the presence of surrounding mountains, while
level Houston has significant chemical production.
As one would expect, when pollution is halted (s = 0), concentration decreases
with time,
c(t) = c o e −t/τ , (6.22)
where c o is the pollution level at the time of stoppage.
6.4 Power Plant Plumes

Pollution dispersal is complicated because of the effects of air turbulence and ther-
mal eddies are more significant to pollution dispersal than the effect of classical
molecular diffusion. Local geography further complicates matters due to its effect
on air currents. Shifting winds reduce pollution levels by spreading out power
plant plumes. A plume shifting sideways by a distance of its width reduces aver-
age concentration by 50%, but over a doubled area. Shifting winds reduce deaths
by reducing some pollution levels below the threshold level at which certain health
problems are triggered. However, the number of deaths is not changed for health
effects that increase linearly with dose. Here we consider only gaseous pollutants;
the transport of 2–10 μ particles in the air must also be considered.
6.4.1 Diffusion Equation

The diffusion equation for a still medium is obtained from Fick’s law and the
continuity equation with Gauss’ theorem. The flux f (kg/m2 s) of an impurity is
proportional to the gradient of its concentration c, as given by Fick’s law,
f = −D∇c, (6.23)
P1: GFZ
6.4. Power Plant Plumes 145
where D is a diffusion constant. Fick’s law is applicable to calculations on diffusion

of impurities in solids, liquids, and gases, and to heat diffusion in solids. If con-
centration is constant, there is no entropic reason for diffusion to produce a flux to
create local minima or maxima.
The volume integral of pollution concentration gives the total quantity of pollution
P in volume V,

P = c dV. (6.24)
The pollution exiting through a surface area A surrounding the volume reduces
total pollution inside the volume. The loss rate of pollution inside the volume is
given by the surface integral of the flux,

S = dP/dt = ∂c/∂t dV = − f · dA. (6.25)
Gauss’ theorem gives

dP/dt = ∂c/∂t dV = − ∇ · f dV, (6.26)
which gives the continuity equation,

∂c/∂t + ∇ · f = 0. (6.27)
The divergence of Fick’s equation with D constant gives
∇ · f = −D∇ 2 c. (6.28)
Inserting the continuity equation gives the diffusion equation when the wind is
absent:
∂c/∂t = D∇ 2 c. (6.29)
6.4.2 Instantaneous Source on an Infinite Plane

A constant area density of pollution p (kg/m2 ) is released all at once on the yz-plane
at x = 0. The pollution spreads in both the +x and −x directions, but not along the
yz plane. This reduces the diffusion equation to one-dimension,
∂c/∂t = D ∂ 2 c/∂ x 2 , (6.30)
which has Gaussian solutions,
c(x, t) = [ p/σ (2π )1/2 ] exp(−x 2 /2σ 2 ) (6.31)
with a Gaussian half-width σ = (2Dt)1/2 . (This is readily verified by substituting
c into the diffusion equation.) At t = 0, c is a delta function with all the pollution
residing in the yz-plane at x = 0. The plume width σ broadens with time (t 1/2 ) and
the peak amplitude at the origin falls with time (t −1/2 ), keeping the total amount of
pollution constant.
P1: GFZ
It would be a large mistake to apply the atomic diffusion coefficients from kinetic
theory of gases to a pollution plume. As an example, perfume from an atomizer is
noticed within a minute on the other side of a room. The travel of perfume mist
is comparable to that of CO2 in air. The standard deviation of CO2 in air after for
1 min with atomic diffusion D = 1.6 × 10−5 m2 /s is
σ = (2Dt)1/2 = (2 × 1.6 × 10−5 m2 /s × 60 s)1/2 = 0.04 m. (6.32)
This value, which is based on atomic diffusion theory, is three orders of mag-
nitude smaller than the experimental value of 10 m. A factor of 103 increase in
σ increases D by a factor of 106 . Macroscopic diffusion, caused by tiny vortices
and turbulence, abruptly moves pockets of air beyond that of molecular diffusion.
Gusting winds further increase this effect and raise empirical D values.
6.4.3 Continuous Point Source in a Uniform Wind

The pollution concentration in solution for an instantaneous point source S with all
three dimensions equivalent (r 2 = x 2 + y2 + z2 ) is
c = [S/σ 3 (2π )3/2 ] exp(−r 2 /(2σ 2 ). (6.33)
3 3/2
A continuous point source [S/σ (2π ) ] is the equivalent of the integral of all past
individual pollution puffs, with the older puffs being spread into wider Gaussians.
The solution for a continuous point source is an error function; however, an easier
solution is available for steady state conditions at distant locations from point
sources such as power plants.
To obtain this solution, Fick’s equation must be modified to include wind velocity
u in the positive x direction. The wind causes an excess flux uc in the direction of
the wind and a diminished flux −uc, against the wind, giving
f = uc − D ∇ c. (6.34)
The divergence of this, with u and D constant, gives the diffusion equation for a
moving medium,
∂c/∂t + u · ∇c = D ∇ 2 c. (6.35)
For continuous pollution (∂c/∂t = 0) at large distances in the direction of the
wind (x 2 y2 + z2 ), this simplifies to
u ∂c/∂ x = Dy ∂ 2 c/∂ y2 + Dz ∂ 2 c/∂z2 , (6.36)
with macro diffusion constants Dy and Dz . The solution is easily shown by substi-
tution to be

c = [S/(2πuσy σz )] exp − y2 /σy2 + z2 /σz2 /2 (6.37)
where S is the pollution emission rate (kg/s) and the plume half-widths are
σy = (2Dy x/u)1/2 and σz = (2Dz x/u)1/2 . (6.38)
P1: GFZ
6.4. Power Plant Plumes 147
6.4.4 Source at 1 km
SO2 concentration depends on three dimensions: the wind’s x-direction from the
source, the horizontal (cross-wind) y-direction and the height of the plume in the z-
direction with respect to the top of the smokestack. Consider the center of a plume at
two locations, 1 km and 10 km downwind from a power plant. The center of the
horizontal plume from a 300-m smokestack lies at z = 0, giving the plume center at
a 1-km distance with coordinates (1,0,0). In the following calculation we assume a
wind velocity ux = 5 m/s and a 1-GWe power plant that burns 104 tons/day of dirty
2% sulfur coal without scrubbers. We assume slightly unstable stability conditions
with macro diffusion constants of Dy = Dz = 25 m2 /s. The plume half-width at
1 km is
σ = σy = σz = (2Dy x/u)1/2 = [(2)(25)(1000)/5]1/2 = 100 m (6.39)
and σ = 316 m at 10 km. The sulfur emission rate S is
S = (0.02 × 104 ton/day)(910 kg/ton)/(8.6 × 106 s/day) = 2.1 kg/s. (6.40)
The SO2 emission rate is 4.2 kg/s since SO2 molecular weight is twice that of
sulfur. This gives an SO2 concentration at a distance 1 km downstream at the center
of the plume,
c(1, 0, 0) = (4.2 kg/s)/(2π )(5 m/s)(100 m)2 e 0 = 13 mg/m3 , (6.41)
or c(1, 0, 0) = 5.1 ppm, using the conversion 2.6 mg/m3 = 1 ppm. The concentra-
tion should also include the SO2 that bounced upward from Earth’s surface in a
reflected plume. Not all pollution bounces, as some is absorbed, but we ignore this.
The additional concentration c’ at (1,0,0) from ground reflections is that of a mirror
image source at z = −600 m, which has negligible value at 1 km, which is
c (1, 0, 0) = (5.1 ppm) exp[−0.5(600 m)2 /(100 m)2 ] = 10−7 ppm. (6.42)
Near the ground at 1 km, the reflected plume is important as it doubles the
concentration: (The coordinate of ground level is z = −300 m = −0.3 km.)
c(1, 0, −0.3) = (2)(5.1 ppm) exp[−0.5(300 m)2 /(100 m)2 ] = 0.11 ppm. (6.43)
6.4.5 Source at 10 km
The SO2 concentration 10 km from the source at plume center and at smokestack
height is c(10, 0, 0). It is one-tenth the c(1, 0, 0) value since σ 2 increases by the ratio
of the distances,
c(10, 0, 0) = c(1, 0, 0)(1 km/10 km) = 5.1 ppm/10 = 0.51 ppm. (6.44)
The reflected plume at 10 km now contributes more significantly, with
c (10, 0, 0) = (0.51 ppm) exp[−0.5(600 m)2 /(316 m)2 ] = 0.08 ppm. (6.45)
P1: GFZ
The sum of the two contributions gives a total concentration in plume center,
c total (10, 0, 0) = c(10, 0, 0) + c (10, 0, 0) = 0.51 ppm + 0.08 ppm = 0.59 ppm.
(6.46)
Concentration in plume center was reduced from 5.1 ppm at 1 km to 0.59 ppm
at 10 km, but the plume is now three times broader with σ = 316 m. Ground level
concentration at 10 km is
c(10, 0, −0.3) = (2)(0.51 ppm) exp[−0.5(300 m)2 /(315 m)2 ] = 0.65 ppm, (6.47)
which exceeds the concentration at the 10-km plume center (0.59 ppm), and it is
6 times larger than ground level pollution at 1 km (0.11 ppm). People who live 10 km
from the plant might agree that the solution to pollution is not dilution, since dilution
combined with distance increases the pollution in this particular situation. If bad
health effects happen only when pollution exceeds threshold levels, then dilution
could be helpful, but if bad health is proportional to concentration, the quote is
accurate.
6.5 Automobile Emissions in the LA Air Basin

In the 1950s air quality in the Los Angeles air basin began worsening due to emis-
sions from cars and trucks. Since then agencies responsible for setting air stan-
dards have had to deal with the ever-increasing number of cars in the basin. Under
the Clean Air Act of 1970, EPA sets the National Ambient Air Quality Standards
(NAAQS) for air basins on hydrocarbons (HC), nitrogen oxides (NOx ), sulfur ox-
ides (SOx ), carbon monoxide (CO), lead, and particulate matter with diameters less
than 2.5–10 μ. EPA was given power to obtain standards under laws that mandated
90% reductions in car emissions. In the 1960s, emission rates for uncontrolled cars
were 10 g/mile for HC, 3.6 g/mile for NOx , and 87 g/mile for CO. Table 6.2 shows
that emission standards dropped dramatically from 1970, with California setting
the pace for lower emissions. Emission standards were mostly unchanged in the
1980s, except for a lowering of NOx to 1.0 g/mile in 1981. California mandated
that ultralow-emission vehicles (ULEV) must comprise 2% of sales by 1997; and
zero-emission vehicles (ZEV) were mandated to make up 10% of sales by 2003,
but market reality has softened these goals in 2003. Technologies enlisted to re-
duce pollution include emission gas recirculation (EGR) valves, high-temperature
platinum/rhodium/palladium catalysts, oxygen sensors and computers to control
engine operation to enhance reactions to minimize pollution,
2NO2 ⇒ N2 + 2O2 and CO + hydrocarbons ⇒ CO2 + H2 O. (6.48)
We estimate the 1973 pollution level in the LA basin using the following:
r 50% of the 7 million inhabitants own cars that are driven 10,000 miles/year.
r 75% of LA pollution comes from cars (86% from mobile sources).
P1: GFZ
6.5. Automobile Emissions in the LA Air Basin 149
Table 6.2. Auto emission standards in grams/mile.

HC CO NOx
Uncontrolled pre-1970 11 80 4.0
1972 3.4 39 3.1
US 1977 1.5 15 2.0
1993 0.41 3.4 1.0
2004a 0.13 1.7 0.07
1973 3.1 28 3.1
CA 1975 0.9 9 2.0
1990 0.24 3.4 0.2
LEVb 1997 0.08 3.4 0.2
ULEVc 1997 0.04 1.7 0.1
ZEVd 2004 0 0 0
Sources: EPA, BTS/DOT and CA Air Resources Board.

a Will cover both cars and light SUVs by 2004 and heavy SUVs by 2008.
b Low-emission vehicles.
c Ultralow emission vehicles were mandated to be 2% of sales by 1997.
d The 10% of sales mandate for zero-emission vehicles is being softened.
r The 1973–74 California standard was 3.1 g/mi (HC), 28 g/mi (CO), and 3.1 g/mi
(NOx ), but we assume a typical car doubles this.
r The LA basin of 3600 km2 has a typical inversion level of 500 m.
The 1973 daily mileage was

(7 × 106 cars/2)(104 mi/year)(1 year/365 days) = 0.96 × 108 mi/day. (6.49)
Since vehicles produce 75% of pollution, total mileage is multiplied by 4/3 to
give an effective 1.3 × 108 mi/day. Assuming older cars emit twice the standard, a
doubled emission rate gives
SHC = (1.3 × 108 mi/day)(6.2 g/mi) = 7.9 × 108 g/day (6.50)
SCO = (1.3 × 108 mi/day)(56 g/mi) = 7.2 × 109 g/day (6.51)
SNOx = (1.3 × 108 mi/day)(6.2 g/mi) = 7.9 × 108 g/day. (6.52)
We assume the pollution is uniformly spread in the LA basin’s box of air with a
volume of area A times inversion height H,
V = AH = (3600 km2 )(500 m) = 1.8 × 1012 m3 . (6.53)

The daily rise in pollution concentration is
dc/dt = S/V, (6.54)
which is the daily source of pollution S divided by basin volumeV. The average
measured maximum monthly pollution level for 1973 are designated as c m in Table 6.3.
The very approximate lifetime of a pollutant in the basin is the time to obtain
maximum concentration, which is the maximum concentration divided by the daily
P1: GFZ
Table 6.3. Pollution in the LA air basin (1973). Daily input

concentration rate (dc/dt), average maximum concentration (c m ),
and time to obtain the maximum concentration (τ ).
HC CO NOx
dc/dt 0.44 mg/m3 -day 4.1 mg/m3 -day 0.44 mg/m3 -day
cm 15 ppm, 10 mg/m3 37 ppm, 45 mg/m 0.42 ppm, 0.85 mg/m3
τ 23 days 11 days 2 days
increase in concentration,
τ = c m /(dc/dt). (6.55)
The table shows that NOx remains in the basin the shortest time (2 days) as
compared to HC (23 days) and CO (11 days). This is consistent with the fact that
NOx is the most reactive of the pollutants as it is depleted during the day by solar-
photochemical reactions, NO2 + γ ⇒ NO + O. However, the released free oxygen
combines with O2 to make ozone, as NOx is depleted. After sunset NOx rises with
a 1-day cycle time, similar to the roughly calculated 2-day buildup time. Smog also
contains hydrocarbon aerosols and SOx .
The LA Air Quality Board has now joined a broader jurisdiction called the South
Coast Region. From 1980 to 1997, the number of vehicle miles traveled in the re-
gion increased 78% as population rose 39%. However, clean air regulations caused
ozone to drop 49%, while NOx remained constant at 0.43 ppm (NAAQS limit of
0.53 ppm) and CO dropped 50% to 17 ppm (NAAQS limit of 9 ppm). However,
there are 80 days a year when ozone exceeds the standard of 0.12 ppm (for one
hour). Generally, pollution conditions have improved in the LA basin, which now
competes with Fresno and Houston for the worst urban air in the country.
6.6 Stratospheric Ozone

The ozone hole was discovered by Joseph Farman along the Antarctic coast in 1983.
Between 1968 and 1991, stratospheric ozone level fell 60%, from 330 to 150 Dobson1
units, while the chlorine level rose from 0.8 ppb to 3 ppb. This is of concern since
high-altitude ozone screens out damaging solar ultraviolet radiation at wave-
lengths between 0.29 and 0.32 μ. Depletion of ozone raises melanoma and non-
melanoma cancer death rates by about 4000/year, primarily in people who sun-
burn easily. Peak UV increased by 12% in New Zealand (45◦ S) between 1991 and
1998. Because of these trends, the 160 nations signed the 1987 Montreal Protocol,
followed by the 1990 London revision which banned chlorofluorocarbon (CFC)
1
A Dobson is the ungainly unit of pressure times distance in milliatmosphere-centimeters of
ozone. A Dobson unit multiplied by the number of molecules in 1 cm3 at STP is the number
of ozone molecules above 1 cm2 of the Earth’s surface. One Dobson unit corresponds to 2.7 ×
1020 ozone molecules/m2 .
P1: GFZ
6.6. Stratospheric Ozone 151
production in the developed countries after 1995. Because it is more difficult for
developing nations to respond to a ban on CFCs, the developed nations are re-
quired to assist poorer countries in their shift to replacements. It is projected that
under the protocol the chlorine level will drop from 13 to 3 ppb and the ozone level
will return to normal by 2050. In 2003 it was reported that the ozone-loss rate was
falling
CFC compounds contain chlorine, fluorine, and carbon, such as CFCl3 and
CF2 Cl2 . CFCs do not interact very much as they slowly rise to the stratosphere. In
the stratosphere, CFCs are dissociated by ultraviolet light, releasing chlorine that
attacks the ozone layer. NASA data from 1999 showed that ozone declined 7% per
decade at 10–45 km altitudes during 1979–96. This result is consistent with a 1979
National Academy of Sciences estimate in which NAS predicted 16% reduction of
ozone by 2000.
6.6.1 CFC Reactions

Almost all CFCs arrive in the stratosphere where they are dissociated by solar ultra-
violet rays. The freed chlorine atoms begin a catalytic chain reaction that destroys
ozone by removing one of ozone’s oxygen to make chlorine monoxide, ClO:
Cl + O3 ⇒ ClO + O2 . (6.56)
Next, ClO reacts with another free oxygen, releasing the chlorine to attack another
ozone,
ClO + O ⇒ Cl + O2 . (6.57)
A single CFC molecule annihilates about a thousand O3 molecules before the
chlorine forms a less reactive HCl, which diffuses downward out of the strato-
sphere. Chlorine is also released when HCl interacts with an OH− radical:
HCl + OH− ⇒ H2 O + Cl− . (6.58)
This second process takes place about 50 times before the HCl finally diffuses down-
ward out of the stratosphere. The average number of ozone molecules destroyed
by each Cl atom from CFC is the product of the following branching ratios:
(1000 O3 /CFC)(50 Cl rebirth via HCl) = 50,000 O3 /CFC. (6.59)
6.6.2 UV Production and Annihilation

Before CFCs were introduced, the rate of destruction of O3 by NOx and solar UV
was equal to the production rate by UV. Ozone is depleted by NO from fossil fuel
combustion and supersonic transport emissions. The reaction is
NO + O3 ⇒ NO2 + O2 . (6.60)
Solar ultraviolet rays not only remove ozone from the atmosphere by dissociating
CFCs, but they also create O3 from O2 at the rate of about Isolar = 2 × 1039 O3 /year.
P1: GFZ
Two-dimensional (r, θ ) calculations use variable ozone densities that increase near
the poles and depend on solar flux, CFC sources, reaction rates, and diffusion
coefficients. The results give space (r, θ ) and time distribution functions for the
relevant compounds (CFC, O2 , O3 , NOx , OH, HCl).
6.6.3 Stratospheric Box Model

We use a simple, steady-state, zero-dimensional box to estimate ozone production
and destruction without the Montreal–London Protocols. Global CFC production
in the 1980s was about 1 million tons, equally divided between CF2 Cl2 and CFCl3 .
We assume equilibrium conditions are approximately established after 50 years of
constant production, which is the lifetime of a CFC molecule. The O3 layer extends
from an altitude of 20 km to 40 km with an average density of 3.5 × 1018 /m3 . The
number of O3 molecules per unit area is about
(3.5 × 1018 /m3 )(20 km)(103 m/km) = 7 × 1022 O3 molecules/m2 . (6.61)

In Dobson units this is
(7 × 1022 O3 /m2 )/(2.7 × 1020 O3 /m2 ) = 260 Dobson units, (6.62)
which is close to the average of 300 Dobson units. The number N of O3 molecules
in the stratosphere is obtained by multiplying 7 × 1022 O3 molecules/m2 by the
Earth’s area,
N = (5.1 × 1014 m2 )(7 × 1022 O3 /m2 ) = 3.6 × 1037 O3 molecules. (6.63)
Since this is 2% of the production of O3 by UV (Isolar = 2 × 1039 O3
molecules/year), the ozone layer is replenished many times during a year, with
fluctuations following the solar variations. A production of 0.5 Mton/year CF2 Cl2
(molecular weight of 121) creates (5 × 1011 g/121 g) = 4.1 × 109 CFC moles/year.
This gives
(4.1 × 109 mole/year)(2 Cl/molecule)(6.023 × 1023 molecule/mole)
= 4.9 × 1033 Cl atom/year. (6.64)
A production of 0.5 Mton per year of CFCl3 with a molecular weight of 137.5
creates
(5 × 1011 g/137.5 g)(3 Cl)(6.023 × 1023 ) = 6.6 × 1033 Cl atoms/year, (6.65)
for a total of 1.15 × 1034 Cl/year that arrive in the stratosphere. Each Cl destroys
50,000 O3 , giving an annual ozone destruction in the stratosphere of
(50,000 O3 /Cl)(1.15 × 1034 Cl/year) = 5.7 × 1038 O3 /year. (6.66)
The depletion of the ozone layer in a stratospheric box is determined from a
steady state rate equation,
dN/dt = Isolar − Rozone N = 0, (6.67)
P1: GFZ
6.7. Purifying Water 153
where N is the number of ozone molecules in the stratosphere. The net production
rate of ozone is dN/dt, which is zero for the steady state. The solar UV production
rate of O3 is Isolar = 2 × 1039 /year and Rozone is a rate constant for destruction of O3 .
Before CFCs were introduced, the rate of destruction of O3 by NOx and solar UV was
equal to the production rate by UV. The CFCs destroy an additional 0.57 × 1039 O3 ,
to which must be added the solar UV and NOx destruction, for a total destruction
rate of
UV/NOx + CFC = 2 × 1039 + 5.7 × 1038 = 2.57 × 1039 O3 /year. (6.68)
CFCs increase the initial destruction rate constant Ro by 30% to a final value,
Rf = (2.6 × 1039 /2.0 × 1039 )Ro = 1.3Ro . (6.69)
Using the higher destruction rate Rf gives the final number of ozone molecules
in the stratosphere,
Nf = Isolar /Rf = Isolar /1.3Ro = (Isolar Ro )/1.3 = 0.8No . (6.70)
The 20% reduction in ozone is similar to more sophisticated predictions. Strato-
spheric nuclear explosions can also affect ozone, but this effect is no longer an issue
because the Unite States and the Soviet Union agreed in 1963 to ban weapons tests
above the ground. The number of NO molecules created from 1% of the energy of
a 1-Mton bomb (that is, 10 kton fully) at 2.5 eV/NO is considerable:
(10 kton)(4.2 × 1012 J/kton)(1 eV/1.6 × 10−19 J)(1 NO/2.5 eV) = 2.6 × 1032 NO.
(6.71)
6.7 Purifying Water

A 2003 National Research Council report states that 2 million tons of waste are
dumped into the world’s rivers, lakes, and streams each year. They further suggest
that by 2023 the amount of clean water available will drop by one-third and that
by midcentury 2–7 billion people out of a population of 9.3 billion will experience
water scarcity. Most developed nations have sufficient clean water or the wealth to
purify polluted water, which is not the case in many developing countries.
The United States maintains high quality water by enforcing drinking water
standards.2 Agriculture extensively uses underground aquifers, in many cases de-
stroying long-term water resources. Wealthy countries can import food, an option
that is difficult for poorer nations. The United States has a considerable amount
of renewable water, some 10,000 m3 /person-year, but the southwest, the Ogallala
2
Primary Drinking Water Standards maximum contaminant level (MCL in milligram/liter)
to prevent adverse health effects. Arsenic (0.01), chlorine (0.8), chromium (0.1), copper (1.3),
cyanide (0.2), fluoride (4), lead (0.015) mercury (0.002), nitrate (10) nitrite (1), benzene (0.005),
chlordane (0.002), asbestos (7 million fibers/liter).
P1: GFZ
Table 6.4. Global water balance. Precipitation and

evaporation in km3 /yr.
Precipitation (P) Evaporation (E) Net (P − E)
Ocean 390,000 430,000 −40,000
Land 110,000 70,000 +40,000
Aquifer in the Plains states and the east coast have had shortages, and these will
get worse.
The water have-not nations, such as in Israel (450 m3 /person year), Jordan (300),
and Kuwait (75), are severely threatened. Israeli–Palestinian conflicts were not
caused by water supplies, but it is a factor to be overcome if there will be a final
solution. Indeed, in 1991 Boutros Boutros-Ghali, UN Secretary General, pointed
out water’s role in Middle Eastern conflicts by stating, “The next war in the Middle
East will be over water, not politics.” Though an overstatement, he was correct that
water is an important aspect. Poor nations with water shortages have few good
options. Their citizens can use bad water and have bad health, they can overpump
their aquifers, they can spend money for desalinization, or they can go to war.
Demand for water is exacerbated by growth of irrigation, which accounts for
two-thirds of freshwater use. About 40% of food is grown on irrigated soil and 20%
of this water comes from depleting aquifers. To some extent, this shortfall can be
mitigated by drip-irrigation, by avoiding costly furrow irrigation, and by turning
to diets with less meat.
Humans used 54% of runoff water in 1966. Renewable water supply is deter-
mined by the solar-evaporation cycle that accounts for precipitation and evapora-
tion from land and sea. The land gains 40,000 km3 /year water from excess ocean
evaporation, which returns to the oceans as runoff. The water balance in Table 6.4
is changed by agriculture, logging practices, and climate change.
How much water is enough? The United States with abundant water supplies
uses more than other countries. An American family of four uses 300 gal/day
with 135 gal for toilet use, 100 for bathing, 35 for laundry, and 30 for kitchen. On
a yearly basis, this is 110,000 gal (410 m3 , 0.34 acre feet), which is tripled if one
were to include lawn/garden to 350,000 gal (1300 m3 , 1.1 acre feet). In addition, we
must include water for irrigation and industry, which increases family use by some
300,000 gal/year (1000 m3 /year).
6.7.1 Ultraviolet Purification

Many parts of Asia, Africa, and Latin America use untreated water, a practice
so widespread that it annually kills 2 million children under the age of five from
diarrheal diseases. This tragedy could be overcome if inexpensive technology to de-
stroy bacteria became available to these regions. A. Gadgil of the Lawrence Berkeley
National Laboratory developed an approach using mercury fluorescent tubes with-
out phosphors that allows the copious ultraviolet photons to escape the tube and
P1: GFZ
6.7. Purifying Water 155
destroy bacterial DNA. The total cost for bacteria-free water is 4 /

c per ton (requiring
0.1 kWh/ton), enough to supply 100 people with 10 l/day at 0.04 / c /person-day.
The number of photons produced from a kilowatthour with a mercury fluorescent
tube (4.9 eV at 80% efficiency) is
(0.8)(3.6 MJ/kWh)(1 eV/1.6 × 10−19 J)(γ /4.9 eV) = 3.7 × 1024 γ /kWh. (6.72)
This abundant production is about Avogadro’s number of photons, 6 × 10 pho- 23
tons, per penny. The cost of a unit that purifies 15 l/min is $2000 to $4000. Electricity
in the developing world can be expensive or even unavailable, but photovoltaic
and wind generators are cost effective at an additional cost of $1500.
A florescent dose of 0.02 J/cm2 (20 milliWatt s/cm2 ) is recommended to sterilize
water, but commercial units use 4 times this amount (0.08 J/cm2 ) to destroy 99.995%
of waterborne bacteria and viruses. Commercial units prefilter water, reducing
turbidity, removing chemical impurities, and enhancing UV sterilization. A small
commercial unit sterilizes 4 gal/min with a 40-W tube over an area of 2700 cm2 . At
50% efficiency, UV flux is 20 W/2700 cm2 = 0.007 W/cm2 . Four gal/min provides
a 12-s irradiation period, depositing 0.08 J/cm2 (12 s × 0.007 W/cm2 ).
6.7.2 Distillation
Distillation is the vaporization and condensation of water to remove chemical im-
purities, bacteria, and viruses. Current cost is about $4 per ton, 200 times more
expensive than UV purification, which does not remove chemicals. The thermal
energy to heat (cT) and vaporize (L v ) a ton of water is
cT + L v = (4.2 MJ/◦ C)(80◦ C) + (2400 MJ) = 2800 MJthermal /tonne. (6.73)
The UV approach uses much less energy (0.1 kWhe /tonne = 0.4 MJe /tonne),
which at 1/3 electrical efficiency amounts to 1 MJt /ton. The energy ratio of 2500
exaggerates the UV advantage since heat recovery can retain much of distillation
input energy. New distillation systems are projected to cost $0.50/ton, or 25 times
the UV cost.
6.7.3 Reverse Osmosis

Normal osmosis is the flow of a water solvent through a membrane from a region
of lower impurity concentration to a region of higher impurity concentration. The
reason a person becomes dehydrated from drinking ocean water is that osmosis
causes an osmotic pressure difference (p) to push water from the less saline body
fluids to the saltier stomach, which contains ocean water. The pressure difference
between the two regions is not caused by statistical mechanics of particle concen-
trations; but rather, it is the result of the attraction between the Na+ and Cl− ions.
This attraction lowers the saline solution chemical potential by μ = 0.0004 eV
per molecule of H2 O. The reduced pressure on the saline side of the membrane
forces water to cross the membrane from the purer side. The change in pressure
from osmotic pressure p is the average potential energy difference (3.5% salt in
P1: GFZ
the oceans) of a water molecule divided by the volume of an H2 O molecule, or

p = μ/V = (0.0004 eV)(1.6 × 10−19 J/eV)/(3 × 10−29 m3 ) ≈ 20 atm. (6.74)
In reverse osmosis, external pressure of about 60 atmospheres is applied to the saline
solution to raise the pressure above osmotic 20-atmosphere pressure, reversing flow
by forcing water to the freshwater side. Reverse osmosis is widely used for water
purification in such places as Tampa, Florida (35 million m3 /day), and the Middle
East. The cost has dropped to $0.55/ton, with plants making pure water at the rate
of 60,000 m3 /day.
In principle, reverse-osmosis pressure could be obtained for free by placing a
pipe with a membrane 600 m below the ocean’s surface (ρgh = 103 × 10 × 600 =
6 × 106 Pa), but such a method is not practical. Free power could also, in principle,
be obtained from osmotic pressure from a flow of freshwater into saltwater. For the
case of the Hudson River flowing into the Atlantic, the osmotic power is
P = force × velocity = (20 × 105 Pa)(600 m3 /s) = 1.2 GW. (6.75)
But this is an unlikely source of power since the flow stopped by the membrane
would destroy the membrane.
6.8 Environmental Chemistry

In the discussion on acid rain, we assumed that all the sulfuric acid was doubly
ionized. This assumption is not quite valid, since not all the sulfates will be doubly
ionized. Two outcomes are possible beyond a small amount remaining as H2 SO4 :
H2 SO4 ⇒ 2H+ + SO−2
4 and H2 SO4 ⇒ H+ + HSO−1
4 . (6.76)
The ratio of the two outcomes results from the law of mass action, which requires
the dissociation rate at equilibrium (molecules to ions) be equal to the association
rate (ions to molecules). The dissociation constant 10− pK depends on the energy
difference between the two sides of the equation and the solution temperature.
When species A and B are added to water, some of A and B continue to exist along
side the new species C and D,
A + B ⇐⇒ C + D. (6.77)
Equilibrium is reached when
[C][D] = 10− pK [A][B], (6.78)
where the brackets [ ] represent ion concentration in units of moles/liter. The reac-
tion dissociation constant 10− pK is usually given for 25◦ C temperature.
The most important case is the dissociation of water into ions, which has an
equilibrium concentration,
[H+ ][OH− ] = 10−14 . (6.79)
P1: GFZ
6.8. Environmental Chemistry 157
Pure water without acid or base has equal ionic concentrations of hydronium
H3 O+ (which we indicate as H+ ) and hydroxide OH− . That is [H+ ] = [OH− ] = [x].
Since water is mostly unionized, the number of moles of ions/liter is much less
than it is for water, giving [x]
1. This gives
[x][x] = 10−14 , (6.80)
or [x] = 10−7 with pH = −log10 [10−7 ] = 7. If added acid gives hydronium concen-
tration of 10−6 moles/l with pH = 6, the hydroxide concentration drops from 10−7
to 10−8 . (Note that pH + pOH = 14.)
6.8.1 Monovalent Acids

Things are complicated when pollutant P moles of nitric acid (HNO3 ) is added to
1 l of water. The resulting dissociation produces gives four unknowns: [H+ ], [OH− ],
[HNO3 ], and [NO− 3 ]. The four equations to determine the four concentrations are
as follows:
water : [H+ ][OH− ] = 10−14 (6.81)
nitric acid : [H+ ][NO−
3 ] = 10 [HNO3 ]
1.0
(6.82)
NO−
3 : [NO− 3 ] + [HNO3 ] = P (6.83)
charge conservation : [H+ ] = [OH− ] + [NO−
3 ]. (6.84)
6.8.2 Divalent Acids

When sulfuric acid is placed in water, some sulfates are singly ionized and some
are doubly ionized. The removal of the first hydrogen ion is relatively easy. The
removal of the second proton is more difficult, as it is more tightly bound with a
lower dissociation constant. The five equations for determining the five unknowns
are as follows:
1st H+ removal : [H+ ][HSO−
4 ] = 10 [H2 SO4 ]
3.0
(6.85)
2 nd +
H removal : [H ][SO−2
+
4 ] = 10 −1.9
[HSO−
4] (6.86)
+ − −14
water : [H ][OH ] = 10 (6.87)
SO−2
4 : [SO−2 −
4 ] + [HSO4 ] + [H2 SO4 ] = P. (6.88)
charge conservation : [H+ ] = [OH− ] + [SO− −2 ∗
4 ] + 2[SO4 ]. (6.89)
∗
[SO−2
4 ] is multiplied by 2 because it is divalent.
6.8.3 Air/Water/Solid Interface

Carbon dioxide is absorbed by the oceans, forming carbonate ions. Carbonate ions
combine with calcium ions, precipitating carbon as CaCO3 . This process is de-
scribed by six equations, which can determine the six unknowns; CO2 in solution
p(CO2 ), H2 CO3 , CO− −2 +
3 , CO3 , H , CaCO3 .
P1: GFZ
Henry’s law gives the ratio of the concentration of CO2 in aqueous solution to its
atmospheric concentration (now at 360 ppm). Henry’s law and five other equations
are as follows:
Henry : [H2 CO3 ] = p(CO2 ) KH = (360 ppm)(10−1.47 ) = 10−4.91
(6.90)
st
1 H : +
[H ][HCO−
+
3 ] = 10
−6.35
[H2 CO3 ] (6.91)
−2
2nd +
H : +
[H ][CO3 ] = 10 −10.33
[HCO− 3] (6.92)
+ − −14
water : [H ][OH ] = 10 (6.93)
Ca concentration : [Ca+2 ] = as measured (6.94)
charge conservation : + +2
[H ] + 2[Ca ] = [OH ] + −
[HCO−
3] + 2[CO−2
3 ] (6.95)
6.9 Flowing Water

6.9.1 Steady State Flow
Water enters one end of a lake and exits at the other end. The water flow rate F is the
cross-sectional area Atimes the flux of water f , or F = fA in m3 /s. The pollution flow
rate is the product of water flow rate F times pollution concentration c (grams/m3 ),
or cF in grams/s. All parameters are function of time. If there is no evaporation
or if no additional water enters the lake, the flow rate of water at the two ends of
the lake is the same, Fin = Fout . If the flow rate of pollution entering the lake is the
same as that departing the lake,
c in Fin = c out Fout . (6.96)
For the case of equal flow rates, Fin = Fout , the input pollutant concentration is the
same as that exiting the lake, or c in = c out ,
If 10% of flowing water evaporates (Fout = 0.9Fin ), exiting pollution concentra-
tion rises 10%:
c out = c in (Fin /Fout ) = c in (Fin /0.9Fin ) = 1.1c in . (6.97)
6.9.2 Nonequilibrium Flow

Time dependence of water pollution can be seen in the following scenario. Pollution
enters a lake at a rate c in Fin at a time (t = 0) when there is no pollution in the lake
[c(0) = 0]. Since pollution leaves the lake at a rate of cF when there is no evaporation
(Fin = Fout = F ), the rate of increase in the total quantity P of pollution in the lake is
dP/dt = c in F − cF . (6.98)
The quantity of pollution in the lake is P = cV for a uniform concentration in the
lake volume V. Rewriting the rate equation gives
dP/dt = Vdc/dt = c in F − cF , (6.99)
P1: GFZ
6.9. Flowing Water 159
which has a solution,
c(t) = c in [1 − e −t F /V ] = c in [1 − e −t/τ ]. (6.100)
The ratio of the lake’s volume to its flow rate, τ = V/F , is the replacement time
constant τ , as well as the exponential time constant. It is large for large volumes and
low flow rates. The lake concentration c asymptotically approaches the incoming
concentration level c in for t τ .
6.9.3 Lake Woebegone

Consider a fictional Minnesota lake with 10 km2 area and 10 m depth. Lake
Woebegone was pristine until a factory started dumping 100 tonnes/year nitric
acid into the lake’s entering stream at a flow of 1 m3 /s. The incoming pollution
concentration was
c in = (dP/dt)/Fin = (108 g/year)/(1 m3 /s)(3.1 × 107 s/year) = 3.2 g/m3 . (6.101)
This corresponded to a molar concentration of
[3.2 g/63 g HHO3 mole] = 0.05 moles/m3 = 5 × 10−5 moles HHO3 /l, (6.102)
which has a pH of
pH = − log[5 × 10−5 ] = 4.3. (6.103)
The exponential time constant is
τ = V/F = (108 m3 )/(3.1 × 107 m3 /year) = 3.2 years. (6.104)
The company closed its doors after 2 years with a level of
c o = c in (1 − e −2/3.2 ) = (3.2 g/m3 )(1 − 0.54) = 1.5 g/m3 . (6.105)
At that point, the pollution level decreased at a rate
dc/dt = −cF/V, (6.106)
which has the solution
c(t) = c o e −t/τ . (6.107)
Six years after the plant closed, the pollution level in the lake dropped to
c(6 years) = (1.5 g/m3 )(e −2 ) = 0.2 g/m3 , (6.108)
with a pH of 5.5, the level for rain from pristine air with 360 ppm CO2 . This result ig-
nores nitrate reactions with biological and mineral materials, which precipitate the
pollution to the lake bottom, lowering the concentration. This model also ignores
evaporation, gravitational stratification, and incomplete mixing.
P1: GFZ
Problems
6.1 Acid rain. The density of sulfur emissions in the east United States is now
about 10 tons per square mile per year. What is the equivalent pH of acid rain
in a dry year with 0.5 m of rain? In a wet year of 1.2-m rain? Assume SO2
accounts for two-thirds of the acid in rain.
6.2 1979 SO2 . EPA relaxed scrubbing of western coal from 90% to 70% sulfur
removal, unifying the emission rates in sulfur/kWh across the country. If
eastern coal contains 1.5% sulfur by weight and it is scrubbed at a rate of 90%,
what is the likely sulfur content of western coal?
6.3 1990 SO2 . EPA mandated a reduction of 10 Mton of SO2 emissions, a drop
of 40% over previous emissions. Assume United States has 285 GWe of coal
plants operating at 35% efficiency with a load factor of 60%, and coal with
11,000 Btu/pound with 1.5% sulfur. (a) What is the amount of sulfur emis-
sions without scrubbers? (b) How many plants need to be equipped with 90%
removal scrubbers to obtain a reduction of 10 Mtons? (c) Will running the
plants with western coal having 0.7% sulfur, scrubbed at 70%, satisfy EPA?
What are emissions if western coal is not scrubbed?
6.4 Pollution scaling. (a) Determine pollution scaling a square city (Section 6.3)
with the wind blowing along the diagonal. (b) What is average pollution rise
time for a square city when wind is blowing along the diagonal?
6.5 Lead scaling. Lead levels in US children dropped 90% after lead was removed
from petroleum. In Third World cities, lead levels in blood were measured at
120 μg/l for cities of 10 million population, at 50 μg/l for cities of 1 million, and
50 μg/l for cities of 0.1 million. Does a scaling apply to this situation? In these
cities, gasoline contained 0.2 g Pb/l. Using reasonable estimates, determine
coupling between lead in air and lead in children’s blood (Thomas, et al.,
1999).
6.6 mg/m3 to ppm. Show that 2.6 mg/m3 of SO2 is equivalent to 1 ppm by volume.
6.7 Transient scaling. Consider a 30 km × 30 km city of 2 million. Citizens con-
sume electricity at an average of 1.5 kW with a sulfur emission rate of 2.5 lb
SO2 per 100 kWh. Industrial pollution raises this sulfur rate by an additional
30%. Assume 70% of pollution is emitted during the daytime. Wind velocity
is 5 m/s in a direction parallel to a side of the square, and the inversion height
is 500 m. What is the rise time for the transient portion of pollution? Plot the
pollution level of the steady state oscillation of SO2 in ppm.
6.8 20 GW in a box. Twenty GWe are generated in a western state. (a) At 38%
efficiency, how many tons of coal with 12,000 Btu/pound are burned in a
year? (b) How many tons of SO2 and CO2 are released per year from 1% sulfur
coal? (c) The state has a 500 km × 500 km area and its inversion height is
1 km. What is the concentration of SO2 after one day’s emissions? How does
it compare to the 24-h, 0.14 ppm standard?
6.9 Instantaneous plume in one and three dimensions without wind. Show that
Gaussian solutions (Eqs. 6.31 and 6.33) satisfy the diffusion equation for the
case of no wind (Eq. 6.30) in one and three dimensions. A puff of 1 kg of SO2
P1: GFZ
Problems 161
is emitted without wind with a diffusion constant of 25 m2 /s. What is the SO2
density at a radius of 100 m after 10 min?
6.10 Continuous plume in three dimensions with wind. Show that the Gaussian
solution (Eq. 6.37) is a solution to the diffusion equation with wind (Eq. 6.36)
for the case of x 2 y2 + z2 .
6.11 San Luis air basin. Twenty thousand cars drive 20 miles/day in San Luis
Obispo air basin, which measures 20 km × 10 km, a region surrounded by
mountains. If cars emit the 1990 NOx standard of 0.4 g/mi, what is the daily
NOx concentration for a 500-m inversion height?
6.12 Ozone. A Dobson unit is a milliatmosphere centimeter ozone (pressure times
distance unit). Show that a Dobson unit is 2.7 × 1020 ozone molecules per
square meter.
6.13 Acre feet. It takes about 1.0 acre feet of water a year to sustain an American
family of four in all of its aspects. How many gallons/person-day is this?
6.14 Reverse osmosis. Estimate the size of a water molecule needed to obtain the
osmotic pressure of 23 atmospheres. Design a power plant that uses osmotic
pressure.
6.15 Polluted water. The limit on nitrates in drinking water is 10 mg/l. Use the
four equations in Section 6.8 to determine the concentrations of [H+ ], [OH− ],
[HNO3 ], and [NO− 3 ].
6.16 Great Lakes flow. The HOMES ratio of the volumes of the Great Lakes
is 7/3/10/1/25 = Huron/Ontario/Michigan/Erie/Superior. The volume of
Lake Erie is 25,000 km2 × 20 m and the flow rate from Michigan and Supe-
rior is 65 km3 /year each. Ignore other flows and evaporation and speculate
that Chicago puts 100 tons/year of nitric acid into Lake Michigan. (a) What is
water-residence time and pollution-rise time for each lake? (b) What is steady
state nitric acid concentration and pH for each lake? (c) Plot pollution concen-
tration as a function of time for each of the lakes, assuming each lake is well
mixed.
6.17 A cleaned lake. A toxic factory emits 10 tons/day nitric acid into Lake Fuzzy
Logic, which has a volume of 10 km × 10 km × 0.1 km. (a) What is steady
state pH and concentration of hydronium? (b) The factory is shut down after
many years of operation; how long does it take for excess hydronium to drop
90%? What is pH at this point in time?
6.18 Coupled lakes with evaporation. Alpha and Beta lakes with volumes Vα and
Vβ are on the Gamma River. Water flows into Alpha at the rate of Fα and
evaporates at rate E α , and then flows into Beta, evaporating at E β . A tributary
adds water to Gamma River, which flows between the lakes at the rate Fγ . A
factory puts pollution into the Alpha at the rate Sα . (a) What is the flow rate out
of Alpha; into Beta; and out of Beta? (b) What is the residence time of water in
Alpha and in Beta? (c) Write the coupled differential equations that describe
the concentration in the two lakes. What assumptions did you make? (d) What
is the steady state level of pollution in the two lakes? (e) Solve the coupled
differential equations and sketch the time dependence of concentrations after
the pollution source is turned on and then turned off at time Toff .
P1: GFZ
Bibliography
Benedick, R. (1998). Ozone Diplomacy, Harvard University Press, Cambridge, MA.
Boeker, E. and R. Grondell (1995). Environmental Physics, Wiley, New York.
Corbitt, R. (1999). Standard Handbook of Environmental Engineering, McGraw-Hill, New York.
Dar Lin, S. (2001). Water and Wastewater Calculations Manual, McGraw-Hill, New York.
Easterbrook, G. (1995). A Moment on the Earth, Viking, New York.
Gadgil, A. (1998). Drinking water in developing countries, Ann. Rev. Energy Environ. 23,
253–286.
Godish, T. (2001). Indoor Environmental Quality, Lewis Press, Boca Raton, FL.
Harte, J. (1988). Consider a Spherical Cow, University Science Books, Sausalito, CA.
———(2000). Consider a Spherical Cylinder, University Science Books, Sausalito, CA.
Harte, J., et al. (1991). Toxics A to Z, University of California Press, Berkeley, CA.
Jain, R., et al. (2002). Environmental Assessment, McGraw-Hill, New York.
Lomborg, B. (2001). The Skeptical Environmentalist, Cambridge Univ. Press, Cambridge, U.K.
Manahan, S. (2001). Fundamentals of Environmental Chemistry, Lewis Press, Boca Raton, FL.
Maidment, D. (Ed.) (1993). Handbook of Hydrology, McGraw-Hill, New York.
National Research Council (1986). Acid Deposition, Long Term Trends, National Academy
Press, Washington, DC.
———(1991). Rethinking the Ozone Problem in Urban and Regional Air Pollution, National
———(1993). Protecting Visibility in National Parks and Wilderness Areas, National Academy
———(2000). Asthma and Indoor Air Exposure, National Academy Press, Washington, DC.
———(2000). Clean Coastal Waters, National Academy Press, Washington, DC.
———(2000). Research Priorities for Airborne Particulate Matter, National Academy Press,
Washington, DC.
Rogner, H. (1997). An assessment of world hydrocarbon resources, Ann. Rev. Energy Environ.
22, 217–262.
Russell, A. (1997). Air quality monitoring, Ann. Rev. Energy Environ. 22, 537–588.
Seinfeld, J. and S. Pandis (1998). Atmospheric Chemistry and Physics, Wiley, New York.
Spiegler, K. (1996). Principles of Desalination, Academic Press, New York.
Thomas, V. et al. (1999). Effects of reducing lead in gasoline, Environ. Sci. Technol. 33, 3942.
Wilson R. and Crouch E. (2001). Risk-Benefit Analysis, Harvard Univ. Press, Cambridge, MA.
P1: GFZ
7
Nuclear Pollution
7.1 Low-Dose Radiation

Radiation has many positive uses, such as medical diagnostics, cancer therapy,
radiosurgery from many collimated beams, nondestructive testing of structures,
smoke alarms, and food irradiation. However, the use of radiation must be con-
trolled because of possible health effects. Without taking time to think about pos-
sible danger, Louis Slotin used his hands to separate a critical mass at Los Alamos
on May 21, 1946. He died in nine days, the first victim of the postwar nuclear
arms race. Slotin did not die of cancer; rather his body stopped functioning after a
dose of over 10 sieverts (1 Sv = 100rem). A one-time dose of 4 to 5 Sv is lethal to
50% of victims. This chapter does not deal with such large doses, but rather with
doses under 10–100 mSv, which can cause cancer after a latency period of some
10–30 years.
The additional death rate is linear at higher doses as seen in the data on Japanese
atomic bomb survivors in Fig. 7.1. Can this linear relationship be extrapolated to
risk at the low-dose region? Or is there a threshold dose below which biological re-
pair is significant? Hiroshima/Nagasaki data show a linear mortality for radiation
dose above 200 mSv (20 rem), but below this threshold the data are more uncer-
tain.1 Data for uranium miners above 100 mSv also point to linearity. Can this linear
effect be extrapolated into low doses of less than 1 mSv? US citizens receive an av-
erage radiation dose of 3.60 mSv/year, most of it—3.0 mSv—from natural sources,
including radon in buildings, and an additional 0.6 mSv from manmade sources,
such as medical x-rays. Natural radiation damages most of the cells in our bodies
every year. Experiments show that the number DNA breaks in cells is proportional
to dose at a rate of 6.3 DNA breaks per human cell per gray (see radiation units
in Section 7.1.1). These experiments show that most of the breaks were repaired
1
The two atomic bombs dropped on Japan had quite different characteristics.
Hiroshima, at 1000 m, gamma ray dose = neutron dose = 3 gray
Nagasaki at 1000 m, gamma ray dose = 10 gray, neutron dose = 1 gray
(Peterson and Abrahamson, 1998).
163
P1: GFZ
164 7. Nuclear Pollution
after a relaxation time of several hours. Figures 7.1 and 7.2 display the 1994 United
Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) data
on Japanese atomic bomb survivors.2 Japanese copper samples are currently being
measured for the content of 63 Ni, which is made from neutron irradiation of copper.
It is planned to separate gamma rays doses (the highest source of radiation) from
fast neutron doses.
7.1.1 Radiation Units

The SI radiation units of gray and sievert are defined below, along with the more
colloquial rem and rad units.
Rate of decay:
1 curie (radiation from 1 g of radium) = 1 Ci = 3.7 × 1010 decay/s
1 bequerel (SI) = 1 Bq = 1 decay/s
Absorbed energy in air:
1 roentgen = 1 R = 87 ergs/g = 0.0087 J/kg
Physical dose of absorbed energy:
1 rad = 100 erg/g = 0.01 J/kg
1 gray (SI) = 1 Gy = 1 J/kg = 100 rad
Biological dose equivalent is absorbed dose times a relative biological effectiveness Q.
X-rays, γ -rays, and electrons have Q = 1, neutrons have Q = 5–20 and alpha
particles and fission fragments have Q = 20.
1 Roentgen equivalent man (rem) = 0.01 J/kg
1 sievert (SI) = 1 Sv = 1 J/kg = 100 rem
Below some threshold, it is not clear from the data whether DNA breaks are suffi-
ciently repaired to prevent the damaged cells from causing cancer. It is very difficult
for epidemiology to settle this issue at very low doses. Some say that the threshold
theory is valid, since the human system does repair DNA after radiation. Others
say that a Taylor expansion of a continuous function of the relationship contains
a linear term. Hence, damage eventually remains. Some say a double DNA break
is needed to cause cancer, and this is continuous with dose. This requirement is
Dose (Sv) Subjects Cancer Expected Excess

< 0.01 42,702 4286 4267 19
0.01–0.1 21,479 2223 2191 32
0.1–0.2 5307 599 574 25
0.2–0.5 5858 759 623 136
0.5–1 2882 418 289 129
1–2 1444 273 140 133
2+ 300 55 23 32
Total 79,972 8613 8106 507
2
Japanese atomic bomb cancer data. Dose to the large intestine (colon) in siverts (Mabuchi,
1998).
P1: GFZ
7.1. Low-Dose Radiation 165
Figure 7.1. Excess relative risk for solid-tumor mortality versus dose for Japanese atomic-
bomb survivors. The error bars correspond to plus or minus one standard deviation. A
straight-line fit to the data yields the high-dose rate relative risk factor of 4.5 × 10−1 /Sv
(4.5 × 10−3 /rem). The two data points below 20 rem are examined further in Fig. 7.2 (Schillaci,
1995).
fulfilled when a natural break and a radiation break occur, supporting the linear
theory. Others say a double break from radiation alone is needed, an idea that
supports a nonlinear theory.
Dose models are key in discussing regulation of nuclear materials in reactors,
wastes, and stored weapons. Low-dose radiation is a predominant issue, as it makes
a considerable contribution in estimates of potential deaths. The Committee on
Biological Effectiveness of Ionizing Radiation (BEIR of the National Academy of
Sciences) accepted the hypothesis that the rate of additional cancers is linear with
dose. However, in 1990 BEIR commented that “at such low doses and dose rates,
it must be acknowledged that the lower limit of the range of uncertainty in the
P1: GFZ
Figure 7.2. Extrapolation of high-dose data to doses below 0.2 Sv (20 rem). The low-dose
data from Fig. 7.1 are compared to possible curve fits based on superlinear, linear, sublinear,
threshold, and hormesis couplings (Schillaci, 1995).
risk estimates extends to zero.” The experts cannot prove the linear theory with
no threshold is correct, but they believe it is correct (2006). A controversial related
issue is that of hormesis, which theorizes that very small radiation doses actually
reduce cancer rates. In Kerala, India, radiation doses are 20 mSv/year, 6 times the
US rate including radon, yet people live longer in Kerala than in the rest of India.
Of course, this longevity might be due to other effects of living in Kerala. The 2006
BEIR report rejected the hormesis theory.
7.1.2 Linear-Quadratic Theory

The absolute risk model, which gives a total probability (Ptotal cancer ) of getting
cancer from radiation and other causes, is described with the relationship,
Ptotal cancer = Pnonrad cancer + Prad cancer = α0 + (α1 D + α2 D2 ) exp(−β1 D − β2 D2 ),

(7.1)
P1: GFZ
7.1. Low-Dose Radiation 167
where α0 is the cancer rate without radiation and α1 is the coefficient for lin-
ear, low-dose radiation D. At higher doses, the quadratic α2 D2 term is impor-
tant, as multiple, closely spaced broken bonds enhance carcinogenic effects. It is
generally believed that double-strand breaks in DNA do not easily self-repair,
and are therefore more dangerous. The double breaks could be caused by one
energetic particle causing multiple damage, or by higher doses. At still higher
doses, the linear-quadratic prediction is diminished, as irradiated cells are de-
stroyed and cannot cause cancer (β1 and β2 ). The natural cancer rate should be
subtracted from the total cancer rate to determine the dependence on radiation.
The usual choice of a linear low-dose radiation considers only the linear term, set-
ting α2 = 0. The BEIR-VII committee rejected (2006) the existence of a threshold
dose level, which would modify D to (D − Dthreshold ), or perhaps take some other
form.
An alternative approach is the relative risk model, which uses radiation dose as a
multiplicative factor to obtain cancer enhancement (Section 9.2). Perhaps the truth
lies somewhere between the absolute risk and relative risk models. Radiation induces
cancer that would not have taken place. It is the hot electrons produced by nuclear
particles that break DNA bonds. Recent studies show that electron energies as low
as 3 eV are sufficient to break these bonds. If DNA double breaks induce cancer,
one can argue that a natural single break plus an ionizing break gives a linear
coupling between radiation and cancer. Smoking and ionizing radiation acting
together have a higher cancer rate by compounding and additive effects, as compared
to acting separately.
7.1.3 Background Radiation

Radiation from natural and human-made sources give an average dose of
3.6 mSv/year (360 mrem/year) in the United States. This gives a death rate of
about 1.5% of the total death rate, under the linear low-dose theory (1990 BEIR-
V). Alternatively, background radiation causes about 7% of all cancer deaths. The
major contributions to background radiation are as follows:
r 2.0 mSv/year (200 mrem) from radon in buildings

r 0.31 mSv/year (31 mrem) from cosmic radiation
r 0.28 mSv/year (28 mrem) from living at sea level
r 0.81 mSv/year (81 mrem) from living in Denver
r 0.39 mSv/year (39 mrem) from natural radioactivity in the body
r 0.3 mSv (30 mrem) from a mammogram
r 0.4 mSv (40 mrem) for a full set of dental x-rays
r 0.1 mSv (10 mrem) from a chest x-ray
r 0.53 mSv (53 mrem) US average annual yearly exposure from x-rays
r 14 mSv (1400 mrem) from a gastrointesinal (upper and lower) x-ray
r 2–9 mSv/year ( 200–900 mrem) for airline flight crews
r 0.05 mSv (5 mrem) for round trip transcontinental flights.
P1: GFZ
7.1.4 Radiation Standards

The limit on annual occupational exposure to radiation dropped from 700 mSv
(70 rem) in 1924 to 300 mSv (30 rem) in 1934, to 50 mSv (5 rem) in 1958, and to
an integrated dose unit of 10 mSv (1 rem) times a person’s age in years in 1993.
The annual limit for continual exposure of the public (nonnuclear employees) was
set at 5 mSv (500 mrem) in 1960 and was lowered to 1 mSv (100 mrem) in 1990.
To set standards scientifically, rule-making authorities need to know the low-dose
coefficients, threshold values, benefit to society from radiation, value of a lost life,
cost to mitigate radiation, and nonradioactive alternatives.
7.1.5 Radiation-Dose Conclusions

The 2006 BEIR-VII/2 report concluded the following on low-dose effects of radia-
tion spread over many persons:
Effects from single incidents of exposure spread over many people:
r BeIR VII: 480 male or 660 female excess deaths per 10,000 person Sv (1 million
person rem), which can be caused by 100,000 persons getting 100 mSv each (10
rem each), or by 10 million persons getting 1 mSv each (100 mrem each). The
average of 570 excess deaths is in a background of 20,000 cancer deaths without
radiation. The 100 mSv dose raises the cancer death rate from 20% to 20.3%. One
statistical death results from 20.8 Sv (2080 rem) for males and 15.2 Sv (1520 rem)
for females. The excess cancer rate is double the mortality rate.
r EPA/NRC (2003) used a risk value of 4 × 10−2 /Sv (4 × 10−4 /rem), which inverts
to 25 Sv (2500 rem) for a statistical death. This gives a 0.8% probability of death
for an exposure of 0.1 Sv (10 rem)
The International Commission on Radiological Protection concluded the follow-
ing in 1991:
r 500 deaths from 10,000 person Sv (1 million person rem)
r 20 Sv (2000 rem) spread over many people = 1 death
r The US background rate of 3.5 mSv/year gives a total dose 0.28 Sv over 80 years.
This dose gives a 1% cancer death rate at 20 Sv/mortality.
r 100 nonfatal cancers per 10,000 person Sv (1 million person rem)
r 130 severe heredity disorders per 10,000 Sv (1 million person rem)
7.1.6 Statistics
How large a population sample is needed to resolve the low-dose issue? About
20% of deaths are caused by cancer.3 A population P has 0.2P deaths due to cancer.
The standard deviation of the number of cancers in a sample of P individuals
3
US annual cancer mortality rate was 200 per 100,000 persons (2 × 10−3 ) during 1978–1988.
US total annual death rate was 870 per 100,000 (8.7 × 10−3 ) during 1980–1998.
P1: GFZ
7.2. Loss-of-Coolant Reactor Accidents 169
is σ = (0.2P)1/2 . How large a sample would it take to observe the effects from
the annual average of medical x-rays of 53 mrem/year (0.53 mSv/year)? Over an
80-year life, medical x-rays give an accumulated dose of 4.2 rem (0.042 Sv), causing
(0.00057 cancer mortality/rem)(4.2 rem)P = 0.0024P, or 0.24% of the population.
If we demand high confidence in our measurements, we require the additional
deaths be at least twice the standard deviation of the measurement. This requires that
0.0024P > 2σ = 2(0.2P)1/2 (7.2)
This condition requires a sample size of P = 0.2 million people who take medical
x-rays and a control group of equal size that does not receive x-rays. A proper
epidemiology study would have to make sure that confounding variables, such
as demography, income, indoor radon, cosmic ray background, air travel, and so
forth have no effect on the result. For this reason, low-dose effects have not been
determined with significant confidence (Section 9.2)
7.1.7 Dose from 10 μCi of 137 Cs

What is the full-body dose to a person if he or she were to put 10 μCi of 137 Cs in a
pocket for a full day? About 50% of the 660-keV (1.1 × 10−13 J) γ -rays is absorbed
by an 80-kg body, and 50% radiates into space. The daily full-body dose for γ -rays
(Q = 1) is
D = (10−5 Ci/2)(3.7 × 1010 /s Ci)(1.1 × 10−13 J)(8.6 × 104 s/day)/(80 kg) (7.3)
= 0.02 mSv/day = 2 mrem/day.
If the source stays in a person’s pocket a year, it would give a dose of 7 mSv/year,
an amount that far exceeds the 1-mSv/year limit on the public for continuous ex-
posure to radiation, as well as the 5-mSv limit on infrequent exposure to the public.
However, 7-mSv dose is legally permissible if the exposure is an occupational risk.
The occupational limit for a nuclear power plant worker, for example, is 50 mSv
(5 rem) for infrequent doses and 10 mSv/year (1 rem/year) for continual doses. If a
radiation source is ingested, most of the gamma rays would be absorbed and dose
would be doubled, with a specific dose to the abdomen rather than to the whole
body.
7.2 Loss-of-Coolant Reactor Accidents

In general, nuclear reactor disasters do not happen because of a single large failure
of safety procedures or technical problem. Rather they happen as a culmination
of multiple, smaller operational failures or mistakes. Reactor malfunction is often
due to poor maintenance practices or minor design errors. A nuclear reactor may
suffer a loss-of-coolant accident (LOCA) from a pipe break that is followed by an
emergency core-coolant system (ECCS) failure. The reactor core will then melt through
the steel reactor vessel, perhaps into the ground below; hence, the nickname for
P1: GFZ
such an accident is the China syndrome. The Browns Ferry accident of 1975 was
caused by a loss of all electrical power, which dangerously evaporated coolant, but
power was restored before the core melted. The 1980 Three Mile Island accident
occurred because cooling water for the core was stopped. The core was partially
melted with the release of small amounts of radioactivity. The heat needed to melt
a core comes from short-lived beta decay of fission fragments, and not from fission
(which has stopped) or from alpha decay of transuranic elements. In this section
we estimate the time for a reactor core to experience a catastrophic failure after a
LOCA.
7.2.1 Reactor Radioactivity

The thermal efficiency η of reactor is 1/3, thus fission develops 3 GWt for 1 GWe
at a fission rate,
(3 × 109 W)(1 fission/2 × 108 eV)(1 eV/1.6 × 10−19 J) = 9.4 × 1019 fissions/s. (7.4)
Each fission event produces two neutron-rich fission fragments, each decaying
about 4 times to reach stability, giving a steady state beta-decay rate 8 times the
fission rate:
dN/dt = 8(9.4 × 1019 /s) = (7.5 × 1020 /s)(1 Ci/3.7 × 1010 /s) = 2 × 1010 Ci. (7.5)
7.2.2 Strontium Production

Strontium-90 is produced in 3% of fission events, annually producing 90 Sr at a rate
NSr-90 = (0.03)(9.4 × l019 /s)(3.2 × 107 s/year) = 9.0 × 1025 90
Sr nuclei/year, (7.6)
90
which corresponds to a Sr mass of
MSr-90 = (NSr-90 )(atomic weight)/NAvogadro
= (9.0 × 1025 90 Sr/year)(90 g/mole)/(6.02 × 1023 90 Sr/mole)
= 14 kg/year. (7.7)
90
Sr has a mean-life τ = 42 years, giving a decay rate
90
dNSr /dt = NSr /τ = (9.0 × 1025 Sr)/(42 year)(3.2 × 107 s/year) = 2 × 106 Ci.
(7.8)
Since September 11, 2001, government agencies have focused on possible terrorist
attacks on reactors and spent fuel ponds. Also of concern is the use of dirty bombs,
made with conventional explosives surrounded with isotopes, such as 137 Cs, 90 Sr,
or 60 Co. A dirty bomb in a large city would not kill a great number of people,
but it would be very expensive to clear the area of the dispersed radioactivity. For
example, there are potential sources of 90 Sr in 300 Russian radiothermal generators
(RTG) that were used to make electricity in remote locations. An RTG can contain
40,000 curies of 90 Sr in a container with a total mass of 1000 kg. Removing 90 Sr from
P1: GFZ
7.2. Loss-of-Coolant Reactor Accidents 171
the RTG produces serious health risks, a fact that serves as a deterrent to the use of
90
Sr in dirty bombs.
7.2.3 Loss-of-Coolant Rise Time

We will calculate the thermal rise time of a light water reactor (LWR) core after
a LOCA. Thermal rise is the time for the core to get sufficiently hot to begin an
exothermal reaction between zircalloy and water. The calculation is based on the
following assumptions:
r Emergency core-coolant water (ECCS) does not arrive until fuel rods are over
1370◦ C, when zircalloy and water exothermically release hydrogen. This happens
below the melting point of 2200◦ C.
r Core mass is 105 kg UO2 for 1 GWe size
r LWR thermal efficiency η = 1/3
r Average fuel temperature is 400◦ C before a LOCA
r Thermal power from beta decay after LOCA (Po = 3 GWt )
P = Po (0.0766t −0.181 ), 0 < t < 150 s (7.9)

P = Po (0.130t −0.283 ), 150 s < t < 4 × l06 s. (7.10)
The temperature rise time is obtained by equating (a) the heat needed to raise
the core to 1370◦ C to (b) the time integral of thermal power P. The heat needed to
raise the core to 1370◦ C is
Q = Nc(T), (7.11)
where N is the number of UO2 moles, c is the UO2 molar specific heat, and T is the
temperature rise for the core to be 1370◦ C, giving T = 1370◦ C − 400◦ C = 970◦ C.
The number of UO2 moles in the core is
N = (108 g)/(238 + 32)g/mole = 3.7 × 105 moles. (7.12)
The high-temperature specific heat, c = 3R = 24.9 J/mole-◦ C, is used since the
temperatures are considerably above the UO2 Debye temperature of 100 K. Thus,
the heat needed to raise the core to its critical temperature is
Qrise = Nc(T) = (3.7 × 105 moles)(24.9 J/mole ◦ C)(970◦ C) = 8.9 × 109 J. (7.13)
The thermal rise time is obtained by equating Qrise to the time integral of the beta
decay power,
t t
Qbeta decay = P dt = 0.0766(3 × 109 )t −0.181 dt = (2.8 × 108 )t 0.819 J = 8.9 × 109 J.
0 0
(7.14)
Solving for t gives a thermal rise time of 68 s, which is close to the published values
of 1 min, calculated with the heat equation (Nero, 1979). Since the time scale for a
LOCA is but a minute, essentially all beta-decay heat is trapped in the core.
P1: GFZ
7.2.4 Loss-of-Power Rise Time

A more gradual LOCA almost happened in 1975 when a workman at the Brown’s
Ferry, Alabama, boiling water reactor (BWR) used a candle to check airflow and
inadvertently set fire to electrical cables, cutting off electrical power for cooling
pumps. Beta-decay heat began evaporating the water coolant, which in turn ini-
tiated a process that would have uncovered the core and begun a LOCA. The
beta-decay heat needed to evaporate 700 tonnes of water is
Qevap = mLevap = (7 × 105 kg)(2.27 MJ/kg) = 1.6 × 1012 J. (7.15)
Setting Qevap equal to the integrated beta-decay heat with two time ranges gives
t = 19 h, similar to the stated 13 h available to recover the situation.
7.2.5 LOCA in Carbon-Moderated Reactors

Reactors that use carbon to moderate neutrons have a longer rise time because of
the higher heat capacity of these reactors. However, the burning carbon moderator
from the 1986 Chernobyl accident propelled radioactivity to great heights and far
distances. Nonetheless, pebble-bed, carbon-moderated reactors could be the basis
of a safer, second-generation of nuclear reactors. Additional thermal mass in smaller
reactors could make LOCAs impossible: There would not be enough integrated heat
to cause damage, but costs would be increased. An improved design might be one
with a smaller core of 0.1 GWe and a lower power density. Raising the heat capacity
of the core with a carbon-moderator could extend LOCA rise times to over 80 h,
hopefully enough time to make repairs. High-temperature gas reactors (HTGR)
use graphite moderators and helium-gas coolant, in contrast to LWRs, which use
water as both moderator and coolant. Helium coolant has the advantage that it
can directly drive a turbine at higher temperatures to obtain efficiencies over 50%,
which is higher than LWR’s 32% using the steam cycle. In addition, new designs
use passive valves and gravity pond pressure to reduce failures.
We calculate the thermal rise time for an HTGR after a LOCA, using the following
assumptions:
r HTGR core contains 500,000 kg of graphite
r HTGR thermal efficiency η = 39%
r Average temperature of the HTGR core is 750◦ C
r HTGR core should be kept below 1700◦ C.
HTGRs have long thermal rise times because their cores have much more heat
capacity than LWRs and they can withstand higher temperatures (1700◦ C vs. 1370◦ C
for LWRs). The HTGR core mass is 5 times greater than the LWR core (5 × 105 kg
vs. 105 kg). In addition, the HTGR graphite moderator has 20 times more specific
heat per unit mass than a LWR since carbon’s mass is 20 times smaller than UO2
(12 vs. 270). The factors of 5 for mass and 20 for specific heat give an HTGR core
100 times larger heat capacity than a LWR. The heat needed to raise an HTGR core
P1: GFZ
7.3. Plume of 137 Cs from a LOCA 173
to 1700◦ C is
Qrise = Nc(T) = (5 × 108 g/12 g/mole)(24.9 J/mole ◦ C)(1700◦ C − 750◦ C)
= 9.9 × 1011 J, (7.16)
which is 140 times greater than the heat we estimated to destroy an LWR. Equating
the integrated beta-decay heat to Qrise gives
150 t
−0.181
Qbeta decay = (10 W/0.39)[
9
0.0766t dt + 0.13t −0.283 dt] (7.17)
0 150
= 109 J[0.46t 0.717 − 2.3] = 9.9 × 1011 J. (7.18)
This heat balance gives t = 12 h, which agrees with more sophisticated calcula-
tions. New designs with smaller modules, lower power density, and ceramic pellets
raise this time to 80 h.
7.2.6 LOCA in Spent Fuel Ponds

The age of terrorism raises the issue of possible attacks on spent fuel ponds. After 1
year, spent fuel radioactive heating is 15 kW/ton and at 10 years it falls to 2 kW/ton.
The spent fuel problem has been exacerbated because density of spent fuel in ponds
was increased as a result of the 1977 decision not to reprocess spent fuel. The
extra fuel rods give additional heating and their presence narrows the infrared and
convection paths to remove heat. Some parameters give temperatures over 900◦ C
after a LOCA, a point where zirconium cladding spontaneously ignites (Alverez
et al., 2003). The problem could be lessened by moving the extra rods to the Yucca
Mountain geological repository or by placing them in surface storage. The damage
could be mitigated by quickly plugging pond holes with quick-setting material,
spraying or pouring water on the ponds or using large air blowers.
7.3 Plume of 137 Cs from a LOCA

A more dangerous, but less likely, result of a LOCA can occur if a reactor’s molten
core comes into contact with enough water to cause a steam explosion. If the explo-
sion is sufficiently large, it might burst the reactor’s concrete dome. Most reactors
have reinforced concrete structures designed to contain radioactivity in the event
of a severe steam explosion accident. Further research seems to have concluded
that it seems likely that the concrete domes would contain the steam explosion, but
some scientists dispute this finding.
7.3.1 Wedge Model Calculation for 137 Cs Dispersal

The number of potential fatalities from an accident is estimated with the wind-
driven diffusion equation (Section 6.4) by calculating deposition from a plume
for each isotope as a function of time. However, there is an easier way to do this
P1: GFZ
problem. The shape of the plume is irrelevant under the assumption of a constant
population density and acceptance of linear low-dose coupling to cancer without
a threshold. For example, if turbulent air broadens the width of the plume by a
factor of two, radioactive concentration is cut in half, but the number of involved
individuals is doubled. Since the number of person-rems in the radioactive wedge is
the same for narrow and wide plumes, the number of cancer mortalities is the same
for both plumes. The American Physical Society’s reactor safety panel obtained
good agreement between the wedge model and results from diffusion equation
models.
We will obtain the dose as a function of distance and then integrate the dose over
the population density to obtain the number of fatalities. The differential volume V
of a wedge-shaped plume is
V = Hθr r, (7.19)
where H is the inversion height, at which the plume stops rising, θ is the wedge
angle, r is the distance from the accident, and r is the radial width of the material
as it moves with the wind. The concentration of radioactivity in the passing plume
(curies/m3 ) is
c = S/V = S/Hθr r, (7.20)
where S is the amount of radioactivity (curies) released by the accident. Most of
the radioactivity, such as 137 Cs, is transported on aerosol particles, not as a gas.
As the wedge plume moves with a wind velocity u, the aerosol particles descend
downward with a deposition velocity as derived from Stokes law,
vdep = ρgd 2 /18η, (7.21)
where ρ is air density, g is acceleration due to gravity, d is particle diameter, and
η is air viscosity. For the case of 2-μ iron particles, the deposition velocity is about
0.2 cm/s. Small particles remain suspended for a longer time, traveling further,
while larger particles settle closer to the reactor.
Ground contamination results from the downward flux of radioactivity from
the shell over the time interval, t = r /u, as it passes overhead. This downward
movement acts like a piston, pushing pollutants downward with the deposition
velocity vdep . If the plume height is 1000 m, it would take 5 × 105 s for the plume to
completely fall to Earth at a deposition velocity of 0.2 cm/s. A wind velocity of 2 m/s
extends the plume to a distance of 1000 km. (Plume radioactivity is considered in
another way in Section 7.5.) Under these assumptions, the ground contamination
g(r ) is obtained from the time integral of the deposition of downward flux:
r/u r/u
g(r ) = cvdep dt = vdep S/Hθr r dt = Svdep /Hθ ru, (7.22)
0 0
with g(r ) in curies/m . Note that the width of the radioactive shell r cancels out
2
since a wider shell gives a lower radioactive density c, but also a longer integration
time. Deposition is multiplied by 4/3 since 137 Cs causes 75% of the long-term
ground contamination. One-half of the 137 Cs inventory (S = 2.9 × 106 curies) is
P1: GFZ
7.3. Plume of 137 Cs from a LOCA 175
assumed to be airborne (Cs melts at 677◦ C) with a wind speed u = 3.5 m/s. The 50%
assumption is probably a worst case estimate. In addition, APS assumed inversion
height H = 1100 m, wedge angle 0.25 radian (14◦ ) and deposition velocity vdep =
0.2 cm/s. These parameters give a ground contamination of
g(r ) = Svdep /Hθ ru = (2.9 × 106 Ci)(0.002 m/s)/(1100 m)(0.25 rad)(r )(3.5 m/s)
= 6000/r, (7.23)
with g(r ) in μCi/m2 and r in km from the reactor. The radiation received depends
on gamma ray energies and shielding from buildings and soil. The APS calculated
a lifetime integrated whole-body dose of 0.0155 sievert for every μCi/m2 of 137 Cs
for persons living their entire lives in the contaminated region. A person’s whole-
body dose is the ground radiation rate g times the dose radiation conversion factor
times a shielding factor of 1/3, for a biological dose equivalent,
Dequiv (r ) = (0.0155Sv/μCi/m2 )(6000/r μCi/m2 )/3 = 31 Sv/r. (7.24)
Therefore, a person living 60 km from the reactor receives a lifetime whole-body
dose of 0.5 Sv (50 rem). At 200 km the dose is 0.16 Sv (16 rem) and at 800 km the
dose is 0.04 Sv (4 rem). The APS used the 1975 value of 0.013 additional cancer
death/Sv absorbed, 10–30 years after exposure (beyond the normal cancer death
rate of 20%). This value inverts to 77 Sv (7700 rem) of low-dose radiation per death.
This is about a factor of 3 higher than 25 Sv (2500 rem) used by EPA and NRC
in the year 2003, (after the 1990 National Research Council’s BEIR-V report) and
about 4 times the BEIR 2006 value of 20 Sv. Respecting the APS report, we present
its calculation, but we increase the result at the end of this section by a factor of 4.
APS determined an increased cancer rate,
Cancer(r ) = (31 Sv/r )(0.013 death/Sv) = 0.4 death/r. (7.25)
A person living at 60 km would have an additional risk of cancer death of 2%
(on top of the normal 20% rate) and a person living at 800 km would have an
additional death risk of about 0.8% (BEIR 2006). It is assumed that people would
leave locations with dose rates higher than 100 μcurie/m2 or at a distance of 60 km
in the APS example. We should increase the death rate total to account for those
that get a very large initial dose.
The number of radiation-related cancer deaths depends on population density
σ downwind from a reactor and each person’s decision to remain or move. The
average population density of the lower 48 states is 30/km2 , with northeast states
at 300/km2 and large cities at a midrange of 3000/km2 . APS used the mid-level
120/km2 to determine the mean number of additional cancers from full-life occu-
pancy. By integrating the plume from the 60-km evacuation radius to 800 km, a
point where most of the radiation has settled out, the mean number of additional
cancer deaths is
r =800
Ncancer = Cancer(r )σ r θ dr = (0.4)(116/km2 )(0.25 rad)(740 km)
r =60
= 8600. (7.26)
P1: GFZ
Since 137 Cs contributes about 75% of the total radioactive dose, the total number
of cancers deaths due to ground contamination is Ncancer = 8600 × (4/3) = 11,500
cancer deaths. The BEIR 2006 factor of 4 raises this to 46,000 cancer deaths. In prac-
tice, this number depends greatly on wind direction and population density, which
is very different for urban and rural locations. Perhaps, the largest uncertainties
lie in the probability of breaking containment and the radioactive fraction that be-
comes air borne, which was small for the Three Mile Island partial core melt but
could be large for a large accident.
To place the above numbers in perspective, population P in the wedge is
r =800
P= σ r θ dr = (116/km2 )(0.25 rad)(8002 − 602 )(km2 )/2 = 9 million. (7.27)
r =60
The integrated population dose is
(11, 500 cancers)(Sv/0.013 cancer) = 880,000 person Sv, (7.28)
or an average of about 0.1 Sv per person for each of the 9 million residents. During 30
years of occupation, the residents also receive a background dose of 3.6 mSv/year
times 30 year for a total of 0.1 Sv, which is the same for the hypothetical accident.
7.3.2 Iodine Contamination

Digested 131 I and 90 Sr from a grass-to-cow-to-milk pathway can be a concern. Io-
dine collects in the thyroid gland, which in turn enhances iodine concentration
by a factor of 7 in adults and a factor of 100 in infants. Most likely, clean milk
would be imported to an affected population after an accident for a period of
weeks to months. Potassium iodide pills can flood the thyroid gland with iodine,
reducing 131 I retention. It took time for the Nuclear Regulatory Commission to con-
vince utilities to distribute KI since the utilities perceived that KI pills on the shelf
would be too worrisome for residents. The relatively short 8-day half-life of 131 I
and the avoidance of contaminated milk can greatly remove the threat of iodine
contamination. However, KI pills will not protect against bone-seeking 90 Sr (T1/2 =
28 year), nor would KI assuage against the effects of dirty bombs, which do not
contain 131 I.
The 1979 Three Mile Island accident released only 20 curies of the core’s 64 million
curies of 131 I. This is not surprising since containment was not breached, but it has
been conjectured that the relatively small release of iodine was a result of iodine
bonding in nonvolatile CsI. At the other extreme, the 1986 Chernobyl accident
released 150 MCi, which was widely dispersed by burning carbon. The estimated
number of fatalities is about 20,000, but this figure has not been documented with
evidence. Chernobyl’s radioactive iodine caused 2000 cases of thyroid cancer in
children under age 14, considerably above the normal rate. While thyroid cancer
has a high cure rate, the effects of radioactive iodine could have been mitigated if
KI pills had been readily available.
P1: GFZ
7.4. Weapon Accident Plutonium Plume 177
7.4 Weapon Accident Plutonium Plume

The threat of plutonium dispersal from a nonnuclear accident with a nuclear war-
head is probably a slight risk, but it has deeper repercussions since it ties into the
nuclear test ban debate. First, we discuss the issue of nonnuclear accidents with war-
heads. The key to safer weapon design is the use of insensitive high explosives (IHEs).
The IHE is less likely to explode with impact, making it less prone to accidental deto-
nation as compared to sensitive high explosives (HEs), which have higher explosive
energy density than IHE. For this reason HE is used to implode the size-constrained
nuclear weapons on MIRVed submarine-launched ballistic missiles (SLBMs). The
rocket fuel on SLBMs is also more energetic and more vulnerable. Intercontinental
ballistic missiles (ICBMs) are not so volume limited, hence IHEs and more resilient
rocket fuel are used on ICBMs. However, this was not always the case. Both ICBMs
and bombers were outfitted with more sensitive explosives until the 1960s. The
shift to safer warheads with IHE and fire-resistant pits was encouraged because of
the Pu dispersal after the 1966 B-52 collision over Palomares, Spain, and the 1968
B-52 fire at Thule, Greenland. The Polomares incident resulted in extensive crop
damage and removal of plutonium-contaminated soil to the United States.
The HE on the Trident W-76 and W-88 warheads allowed designers to maximize
Trident yields at 0.5 Mton for an 8000-km range. With the cold war over, the Trident
SLBMs will be outfitted with only four or five warheads, instead of the previous
eight. The empty space could be filled with larger and safer IHE warheads. But
completely replacing W-76 and W-88 warheads is deemed too expensive by DoD,
as they continue to support the use of HE on SLBMs. Some opponents of the
Comprehensive Test Ban Treaty (CTBT) have called to reestablish nuclear testing
to include testing of new IHE warheads for submarines. Analysis of potential deaths
from nonnuclear warhead accidents is relevant to the CTBT debate.
7.4.1 Wedge Model Calculation for Pu Dispersal

An explosion could be triggered by a bullet shot into SLBM sensitive rocket fuel,
a missile dropped upon loading, or a shipboard fire. All current US implosion
warheads are designed to be “one-point safe.” A warhead needs simultaneous
initiations at two points on the sphere to obtain a useful critical mass. If only
one point explodes, the fissile yield is constrained to exceed four pounds TNT-
equivalent in one out of one million such events. The principal hazard of such
an explosion is Pu-spreading in aerosols to human lungs, causing cancer. In 1990,
Fetter and von Hippel used a wedge model to estimate number of potential fatalities
from Pu released in a nonnuclear warhead explosion. They reviewed the literature
to determine the connection between mortality and dose of weapons-grade Pu
aerosols with worst case droplets with median diameter of 1 micron. They assumed
no protection from the aerosols and determined that of 3 to 12 lung-and-bone
cancer deaths are caused by inhalation of 1 mg of weapons-grade Pu aerosol after
a 10–30 year latency period.
P1: GFZ
We estimate Pu dispersal with the wedge model used for reactor plumes (Section
7.3). The accident mortality rate is the product of the mortality from an accident
times the likelihood of an accident. The combination of linear low-dose coupling
with a constant population density greatly simplifies these estimates. The concen-
tration c of Pu aerosol in air is a function of r , the distance from the accident:
c(r ) = S(r )/(plume volume) = S(r )/Hθr r, (7.29)
where S(r ) is the mass of Pu in aerosol in the air, H is inversion height, and r
is plume width. The Pu trapped in an individual’s lungs is the product of the Pu
concentration c, the breathing rate b in m3 /s, and the time for the plume to pass, t:
Iind (r ) = c(r )bt. (7.30)
The equation for time for the plume to pass is t = r/u, where wind velocity is
u. This simplifies Iind (r ) to
Iind (r ) = [S(r )/Hθr r ] b [r/u] = S(r )b/Hθr u. (7.31)
Note that the moving shell width, r , cancels out. This makes sense, since a wider
shell radius r gives a lower radioactive concentration c and a longer breathing
time. If it is raining, the Pu aerosol is quickly deposited on the ground. If it is not rain-
ing, the quantity of airborne aerosol reduces with distance r as some is deposited
on the ground. Hence, the decay equation for the amount of airborne Pu aerosol is
P(r ) = Po e −r/L , (7.32)
where L is the average distance that an aerosol particle travels before it is
deposited. This distance is
L = uH /vdep , (7.33)
where vdep is deposition velocity. The total amount of plutonium aerosol inhaled
by all people Ipop is determined by integrating the individual amounts over the
affected population:
∞ ∞
Ipop = Iind (r )θr σ dr = [Po e−r/L b/Hθr u] θr σ dr, (7.34)
0 0
where σ is population density. For a constant population density σo , this
integrates to
Ipop = Po bσo L/uH = Po bσo /vdep . (7.35)
As in the case of the Cs plume, wedge angle width θ and height H are not
137
present in the final answer. The parameters used by Fetter and von Hippel give the
amount of inhaled plutonium by the affected population:
Po = 10 kg PuO2
b = 3.3 × 10−4 m3 /s
σo = 30–3000 persons/km2 ; median 300 persons/km2
vdep = 0.003–0.03 m/s; median 0.01 m/s
mortality rate of 3–12 deaths/inhaled mg.
P1: GFZ
7.4. Weapon Accident Plutonium Plume 179
Hence, the highest total amount inhaled is about

Ipop = Po bσ o /vdep = (104 g)(3.3 × 10−4 m3 /s)(3 × 10−4 /m2 )/(0.01 m/s) = 0.1 g,
(7.36)
with mortality of
Ncancer = (100 mg)(3 to 12 deaths/mg) = 300–1200 deaths. (7.37)
The projected number of deaths is now compared to the individual cancer death
rate of about 20%. The plume extends a distance of about twice the mean length,
or
2L = uH/vdep = 2(2 m/s)(1000 m)/(0.01 m/s) = 400 km, (7.38)
for wind speed u = 2 m/s. The number of people irradiated in the wedge is
2L
P= σ r θ dr = (300/km2 )(0.25 rad)(400 km)2 /2 = 6 million. (7.39)
0
A cancer rate of 20% gives a background of about 1 million natural cancers in a

population of 6 million. If the accident caused 600 deaths, it would raise the 20%
cancer death rate by 0.01%, an amount that would be difficult to determine with
epidemiology.
7.4.2 Cost of Weapons Versus Value of Life

Is the loss of 1000 lives, as projected from a worst-case plutonium dispersal accident,
an acceptable risk? There are uncertainties in this estimate, particularly in the amount
of Pu that attaches to aerosols, but we will assume the figure is correct. A true risk
assessment should consider alternative possibilities. In this case, the alternative is
the additional cost of building safer warheads and missiles, as well as the global
political cost of renewed nuclear testing. Here we only compare the value of human
life in contrast to the cost of building new nuclear weapons. If we assume an unlikely
low cost of $2 billion for new SLBM weapons, the carrying cost would be about
$200 million/year. This annual cost amount is equated to probabilistic annual cost
of lost lives:
$200M/year = Ncancer Vlife n. (7.40)
The right side is the product of the number of deaths from an accident (Ncancer ),
the dollar value of human life (Vlife ), and the number of nuclear accident accidents
per year (n). We assign values to Ncancer and Vlife to determine n, the number of
accidents per year to establish financial equality.
We use Ncancer = 1000 and for Vlife we begin with lifetime income of a person.
Assuming 40 years of work at a national average annual income of $35,000 gives
$1.4 million. To the denied income we add the cost of pain and suffering, to arbi-
trarily arrive at V = $5 million. Now the equation is
$200 M/year = (1000 lives)($5 M/life)(n) (7.41)
P1: GFZ
which gives n = 1 accident in 25 years. Thus far no accidents like this have happened.
At $10 million/life, n is reduced to 1 accident per 50 years. Hence the more one
values human life, the smaller is the acceptable rate of nuclear warhead accidents.
In 2003, EPA analysis reduced the value of a life from $6 M to $3.7 M. Further fine
tuning uses the concept of quality of life years. For example, a suffering person that
was saved 4 years through medical mitigation, but still had bad health, would get
credit for 0.5 × 4 year = 2 quality-adjusted life years. Another approach in the
analysis is to value a young life at $6M and an older life at $2million. Also used
is the concept willingness to pay for mitigation, which should be higher for young
people and smaller for older people.
The debate on new IHE warheads arose during debate on nuclear testing prior
to consideration of the CTBT. Those opposed to CTBT said the warheads with
regular HE were not safe enough and further testing of new warheads with IHE
was needed. However, the Pentagon maintained that the HE-loaded warheads
were safe enough and that it was too expensive to rebuild them. On the other
hand, Department of Defense testified in 1992 that a test ban was not a good idea
because DoD wished to continue to test weapons for reliability, safety and new
designs. Legislation required the government to quantify these trade-offs with a
cost versus safety analysis before testing could resume, but this has not been done
since testing stopped. In 2003, DoD asked, and Congress approved, the removal
of the 1993 legislative ban on designing new warheads under 5 kton. Since new
weapons need testing, this would affect the Nuclear Nonproliferation Treaty regime
(Chapter 5).
7.5 Dirty Bombs

7.5.1 Dirty Bombs
The 9/11 attack shows us that terrorists might detonate radioactive sources in
large cities. A dirty bomb attack would probably be more psychological than lethal,
because radiation doses from ground contamination would mostly not be large and
they could be avoided by persons moving away from the area. The 137 Cs, 90 Sr and
210
Po isotopes could be used in dirty bombs. How much radioactivity is needed to
convince a person to leave home? Fear of radioactivity can be illogical. Often the
most educated are the most terrified. It is for this reason that radiological disper-
sal devices (RDD) have been called the “new WMD,” weapons of mass disruption.
The US government is responding by tracking down old sources of radiation on
a global basis, developing alternative technologies to radiation (accelerators for
cancer therapy), improving the detection of smuggled radioactive sources, and
developing better decontamination procedures.
There are thousands radioactive sources that could be used in a terrorist attack:
sterilization (100 to 10 million Curies), radioisotope thermal-electric generators
(10–100,000 Ci), cancer therapy (10,000 Ci), blood irradiators (1000 Ci), radiography
(1–100 Ci), and well logging (up to 100 Ci). One critical issue for a bombmaker is the
P1: GFZ
7.6. Fault Tree Analysis 181
Figure 7.3. Long-term contam-

ination from a cobalt bomb
in New York City. Inner ring:
One cancer death from radia-
tion per 100 persons for those
that remain. Middle Ring: One
cancer death per 1000 people
who remain. Outer Ring: One
cancer death per 10,000 who
remain, a region that EPA rec-
ommends decontamination or
destruction. The Nuclear Regu-
latory Commission might use a
standard 20 times higher, at a
death rate of 1 in 500 for full-
time occupation. In compari-
son, the US background rate of
3.5 mSv/year gives a total dose
of 0.28 Sv over 80 years. This
dose gives a 1 part in 100 can-
cer death rate from natural ra-
dioactivity at a rate; of 20 Sv-
persons per mortality, spread
over many persons. (M. Levi
and H. Kelly, Dirty bombs: Re-
sponse to a threat, Public Interest
Report, Federation of American
Scientists, March/April 2002)
form of the radioactivity material, for example, fine powders and aerosols increase
plume size. Figure 7.3 displays estimates of dirty bomb contamination from a cobalt
pencil used in food irradiation. Clean-up after a bomb explosion would be difficult.
EPA recommends decontamination or destruction within the outer ring where the
death rate is 1 part in 10,000 for continuous occupation. Several bills have been
introduced before Congress to increase controls on radioactive materials, but these
are in conflict with the concerns of medical and industrial users.
7.6 Fault Tree Analysis

Nuclear power has produced 20% of US electricity since the early 1990s, but its
growth has stopped completely. A 1972 projection of 1200 GWe capacity was given
for the year 2000, but this projection fell far short, as actual capacity is 98 GWe .
Global nuclear power capacity remained fairly constant over the past 5 years
at 350 GWe (2000). Seventeen nations use a higher percentage of nuclear power
than the United States, although the United States has the largest capacity. French
P1: GFZ
capacity is 63 GWe , which is 76% of its grid, while Japanese capacity is 43 GWe
for 35% of its grid. Globally 16% of electricity was produced from nuclear power
plants, while 22% comes from renewables (mostly hydroelectric), 63% from fossil
fuels [coal (34%), natural gas (19%), petroleum (10%)]. US electricity (2001) was
produced from coal (53%), nuclear (22%), natural gas (15%), hydro/renewables
(8%), and petroleum (3%).
Increased cost of nuclear power and the 1979 Three Mile Island and 1986
Chernobyl accidents halted plans for new plants. Low cost electricity from nat-
ural gas from combined-cycle gas turbines further removes nuclear power as a
competitor. The continuing unrest about radioactive waste disposal has effectively
forestalled the addition of new nuclear plants.
Under a linear low-dose assumption, the annual number of deaths from nuclear
power is a product of at least eight functions, each of which must first be first
multiplied for each failure mode and then summed over all failure modes. The
eight functions are
r n, number of nuclear plants

r P, annual probability for a failure mode
r S,amount of released radioactivity
r B, biological function, coupling radiation and mortality
r A, plume area
r ρ, population density in the plume
r W, wind and weather function
r t, time spent in irradiated region.
Estimates of an extremely serious reactor accident in a populous location point to

numbers that could be devastating with 3,000 immediate fatalities, 240,000 thyroid
nodules, 45,000 latent cancer fatalities, 30,000 latent genetic effects, and a cost of over
$10 billion, contaminating an area of 8,000 km2 . Three Mile Island was fearful but
not lethal. Nuclear reactors have not had a serious accident and they collectively
have not had a poor record, with total radiation exposure of 5 person-Sv/year,
causing 0.3 of a death per year (Section 7.1). Safety could be enhanced with smaller
cores, lower power densities, and greater heat capacity to totally prevent LOCAs
(Section. 7.2). New reactor designs are not totally passive-safe, but rather they use
passive-safe circulation systems.
Fault tree analysis is widely used by industry to analyze accident scenarios and
estimate relative and absolute accident rates Paccid (Fig. 7.4). Probabilistic risk as-
sessment (PRA) was used to determine Paccid for nuclear power plants in the 1975
Reactor Safety Study under Norman Rasmussen of MIT. This report was criticized
for its underestimation of error bars and for not adequately considering common-
mode failures, such as earthquakes, which can remove more than one safety system
at a time. The study predicted that very serious accidents would be infrequent. Such
accidents could be caused by overlapping small problems. There is consensus that
PRA is useful in determining relative risks, and the Nuclear Regulatory Commission
continues to use this approach.
P1: GFZ
7.6. Fault Tree Analysis 183
(1 – P1)(1 – P2)(1 – P3)

Serial Logic S
(1 – P1)(1 – P2)
(1 – P1)
P
(1 – P1)(1 – P2)P3
P
(1 – P1)(P2)
P
P1
Figure 7.4. Series fault tree for dinner. The upper part is the fault tree for the serial logic
problem of cooking dinner with failure probabilities Pdinner , Pvalve , Ppilot , and Pgas . At path
junctions, the paths split into an upper leg with probability of success S and a lower leg with
probability of failure P with S + P = 1.
7.6.1 Serial Logic

A process made up of many different tasks, each depending on the success of the
previous task, is a problem in serial logic. If there are i different serial tasks, the
probability that the serial process is successfully completed is the product of the
success probabilities for each individual task. Because failure probabilities are small
numbers (Pfailure = ε) and success probabilities are close to unity (S = 1 – Pfailure =
1 – ε), risk analysis uses failure probabilities rather than success probabilities. The
formula describing a system’s success after several serial tasks is
n
Ssystem = 1 − Psystem = (1 − P1 )(1 − P2 ) . . . (1 − Pn ) = i=1
(1 − Pi ). (7.42)
To prepare a meal from uncooked ingredients, we need a stove with a working

valve, a working pilot light, and available nature gas. If all task failures Pi = 10%,
the probability of obtaining a cooked meal is 73%, obtained from
Sdinner = 1 − Pdinner = (1 − Pvalve )(1 − Ppilot )(1 − Pgas ) = (1 − 0.1)3 = 73%, (7.43)
or a probability that cooking dinner fails of 27%. The more tasks involved in a serial
process, the less likely that the process will be successful. See Fig. 7.4.
7.6.2 Parallel Logic

Duplicate back-up systems enhance success by operating in parallel. Reactors have
operating back-up power in case there is a loss of line electricity. A back-up motor-
diesel generator is available when external power fails, and a second generator is
available if the first generator fails. The fault tree diagram in Fig. 7.5 shows branch
points after an electrical power failure. Each branch indicates a chance to recover
with a new, parallel system. The overall probability of maintaining electricity to
operate a reactor depends on the failure probability for line voltage Pline and the
failure probabilities of the two diesel generators, PG1 and PG2 . The path at the
P1: GFZ
Parallel logic
(1 – P1)
S
P1 (1 – P2)
S
P1 P2 (1 – P3)
P1
P1 P2
P
P1 P2 P3
Figure 7.5. Parallel fault tree for reactor operating electric power. Parallel logic describes
back-up power at a nuclear reactor. The failure probabilities are Pelec for reactor operating
electrical power derived from Pline for exterior line voltage and PG1 and PG2 for the reactor’s
motor-generator systems.
bottom of Fig 7.5 is the only one that ends in failure with a probability of
Pelec = Pline PG1 PG2 . (7.44)
The probability for maintaining electrical operating power is the sum of the
successful probabilities for the three task paths,
Selec = (1 − Pline ) + Pline (1 − PG1 ) + Pline PG1 (1 − PG2 ) = 1 − Pline PG1 PG2 . (7.45)
Note that with Selec + Pelec = 1. If each power source has a 10% failure probability,
the system failure rate is
Pelec = (0.1)(0.1)(0.1) = 10−3 or 1 in 1000, (7.46)
which is much better than a failure rate of 27% for a meal prepared by three serial
tasks.
The Rasmussen report estimated a median accident probability of 5 ×
10−5 /reactor-year for a core meltdown, with an upper bound of 3 × 10−4 /reactor-
year (See Table 7.1). The 100 US power reactors, operating for 22 years, have had
2200 reactor-years experience. The pessimistic, highest probability estimate, com-
bined with the US 2200 reactor-years experience, gives a probability for a core melt
Table 7.1. Probabilistic risk assessment. Accident

probability per reactor-year for small, medium, and
large accidents at a light water reactor. [Reactor Safety
Study, Nuclear Regulatory Commission, 1975]
Annual deaths Deaths over 30 years Accident probability
1 30 3 × 10−5 /reactor-yr
100 3000 2 × 10−6 /reactor-yr
1000 30,000 1 × 10−8 /reactor-yr
P1: GFZ
7.7. Geological Repositories 185
accident of
Pcore melt = (3 × 10−4 /reactor-year)(2200 reactor-year/25 year)

= 0.5 accident/25 years, (7.47)
which, by chance, equals the US experience with one-third core melt at Three Mile
Island with a hydrogen bubble. It has often been assumed that 90% of core-melt
accidents would be contained in the concrete domes and that 10% of core melts
would cause a steam explosion that would break containment and spread a plume.
Some claim this assumption is now too pessimistic, based on further research. See
Fig. 7.6 for a complex fault free diagram.
7.7 Geological Repositories

In the year 2000, civilian nuclear power wastes contained 33,000 MCi, which is
slightly larger than military nuclear wastes of 25,000 MCi. Scientists have long en-
couraged geological burial of nuclear wastes, but it is imperative that the repository
effectively contain wastes over long periods of time. The debate for and against
burial hinges on the definitions of effective and risk when compared to alternative
disposition. Debate on waste sites also turns on the issue of “not in my backyard.”
The nuclear waste problem has been badly handled. Liquid waste tanks at
Hanford, Washington, leaked because of a combination of bad chemistry and
single-shell construction. The tanks no longer leak because the radioactive liq-
uids have been removed. The plutonium pit factory at Rocky Flats, Colorado, has
been closed for years because of dispersed plutonium. The former Soviet record is
worse. The dumping of 120 million curies into Lake Karachai is enough to give a
lethal dose to a person standing an hour near the discharge pipe. At another lake
the situation is exacerbated by weather, as drought spreads radioactivity with the
wind, and rains push polluted water over the dam. The Russian government is
now filling the lakes with large stones.
Spent fuel ponds could be terrorist targets, since explosions could make ponds
leak, allowing uncooled rods to melt, burn, and disperse. This could be serious
since the ponds have considerable radioactivity and a uranium fire would propel
radioactivity into a large plume. It might be possible to mitigate such a threat by
spraying water on the pond by plugging the pond holes with fast-drying materials
or by using large fans to blow air. This problem is lessened if cooler rods are moved
into the Yucca Mountain repository or into dry surface storage. This reduces heat
production and increases accessibility of the remaining rods to cooling.
7.7.1 Pu content
A typical 1-GWe LWR discharges about one-third of its 100 ton core every 1.5 years.
Since 0.9% of discharged heavy metal is Pu, the annual discharged Pu from a 1-GWe
P1: GFZ
Figure 7.6. Fault tree of a LOCA from a pipe break. The sequence begins with a failure
probability for a pipe break PA , followed by the failure probability of electrical power PB ,
the failure probability of the emergency core-cooling system PC , the failure probability of the
fission-product removal system PD and the failure probability of the containment structure
PE . Since the probability of success for many of these steps is close to one, this value is
assumed to be unity when that is sufficiently accurate. The reduced tree considers only the
most likely fault paths (Nuclear Regulatory Commission, 1975).
P1: GFZ
reactor is
(100 ton core/3)(0.009 Pu)/(1.5 years) = 200 kg Pu/GWe year. (7.48)
The US 2200 years of operation or power plants produced
(0.2 ton/GWe year)(2200 GWe year) = 400 ton Pu (7.49)
as part of the global total of 1200 tons in 1998.
Reactor-grade plutonium contains over 20% 240 Pu. Nonnuclear weapon states
can make weapons of only about 1 kton yield with reactor-grade plutonium, while
experienced nuclear weapon states can obtain full yield with reactor-grade Pu.
Thus far, the eight nations that have Pu weapons all use reactor-grade Pu for their
weapons. Nevertheless, there is concern that spent fuel containing reactor-grade
plutonium remains in many places without adequate physical security. A PWR fuel
assembly contains 5-kg Pu, enough to construct a nuclear weapon. Its radiation bar-
rier at a distance of 1 m drops from 65 Sv/h after 5 years to 9 Sv/h after 50 years, an
amount sufficient to deter rational actors. A much greater threat is Russia’s 150 tons
of weapons-grade Pu and 1000 tons highly enriched uranium. The United States
and Russia have agreed to work together to dispose of their excess weapons grade
materials. In general, Pu in glass or ceramic logs can be placed into a geological
repository. It can be used as plutonium–uranium mixed oxide (MOX) fuel in reac-
tors; it can be transmuted in accelerators (unlikely because of expense); it can be
shot into space (very unlikely since 1% of rocket launches fail and it is expensive);
it could be deposited in very deep ocean trenches, such as the Challenger Deep,
located in the Marianas Trench. The Challenger Deep is 11-km deep and slowly
subducts into the Earth. This is technically sound but it violates the treaty that bans
ocean dumping.
7.7.2 Funding for Nuclear Waste Disposal

In 1977 the US government offered to accept utility spent fuel for a repository fee of
0.1 c/kWh. The electrical utilities have charged the government with not fulfilling
this agreement. The courts agreed, requiring DOE to pay utilities to keep spent fuel
on-site until such time as Yucca Mountain repository can accept it, perhaps by 2010.
By 2002, the trust fund received some $10 billion and it will have $15 billion by 2010.
If Yucca Mountain cannot accept the fuel, retrievable surface storage is likely to be
used until another repository can be found. In 2004, the US Court of Appeals in DC
ruled against the 10,000 year limit on radiation safety at Yucca Mountain. The court
concluded that EPA must either issue a revised standard that is “consistent with”
the NAS peak-dose standard “or return to Congress and seek legislative authority
to deviate from the NAS report” used until another repository can be found. US
spent fuel in 2000 from 2200 GWe years of operation is about
(20 ton/GWe year)(2200 GWe year) = 42,000 tons, (7.50)
with 80,000 tons expected by 2020, to complete the first generation of nuclear power
plants. Yucca Mountain will have an initial capacity of 77,000 tons (70,000 tons spent
P1: GFZ
fuel plus military waste), which allows it to handle US commercial spent fuel until
about 2015.
7.7.3 Heat Loading

In the first several hundred years, spent fuel beta-decay heat comes primarily
from fission fragments 90 Sr, 90 Y, 137 Cs, and 137 Ba with half-lives of 3 to 30 years.
After 1000 years, heat comes primarily from alpha decay of the actinides 239 Pu,
240
Pu, and 241 Am. After a public rule-making process, EPA used a time horizon
of 10,000 years for regulatory decisions on Yucca Mountain. Some estimates give
doses up to 0.15 mSv/year at 18 km from the repository after 10,000 years, which
was the limit recommended by a National Research Council study. This is less
than the public’s limit of 1 mSv/year. Some estimates show that it could rise to
0.03 Sv/year after 400,000 years, but there is little certainty on this. The discovery
of 36 Cl, a by-product of 1950s testing, indicates water seeps downward at a faster
rate than expected. A regulatory issue is how to link the risks of each step in a mul-
tiplicative process. Should parameters for estimating risk be mid-range estimates
or extreme values? Another issue is whether a rigid radiation standard in a region
with little population is the correct logic. One might think that it is more logical to
compare the number of projected deaths from Yucca Mountain with the projected
number of deaths from other storage choices, since the radioactive material must
be disposed somewhere.
Spent fuel rods can be placed in geological storage at varying heat loads, ranging
from 2 kWt /ton at 10 years to 1 kWt /ton after 30 years. See Table 7.2. Since the
Yucca Mountain site is oxidizing, there is concern that metal containers will be
breached over a period of 1000 to 100,000 years. Because water percolation rates
to the repository are higher than expected, the design uses engineered barriers to
supplement geological barriers. There are very few metals other than gold that
can resist corrosion for such long times, but stainless steel/nickel canisters with
titanium drip shields are believed (but not completely tested) to be robust for
thousands of years. Radioactive heat should keep the canisters above 100◦ C to
keep water away over the initial 1000 to 2000 years.
Table 7.2. Fuel rod heat and radioactivity. The thermal

heat rates (kWt ) and radioactivity rates (Ci) are given for
30 tons of spent fuel, the amount removed from a 1-GWe
reactor every 18 months. [D. Bodansky, Nuclear Energy,
American Institute of Physics Press, 2004]
Age Thermal power Radioactivity
1 year 300 kWt 70 MCi
10 years 70 kWt 14 MCi
100 years 10 kWt 1.4 MCi
1000 years 2 kWt 61 kCi
P1: GFZ
One possible design for a waste package contains 21 fuel assemblies, each having
a 12-ton mass with 12 kWt heat at 30 years of age. To simplify matters, we assume
the cylindrical package is a sphere for distances larger than its size. The decay
heat power Pheat from the “spherical” container in thermal equilibrium causes a
temperature gradient T between spherical shells of thickness r . This allows us
to write
Pheat = −k4πr 2 T/r, (7.51)
where r is radial distance of the shell and k is thermal conductivity of the geological
media. Letting T/r = dT/dr and integrating r from the canister surface to a
distant point, we obtain the temperature T(r ) as a function of radial distance r ,
T = Tsurface − T(r ) = (Pheat /4π k)(1/rsurface − 1/r ). (7.52)
The temperature rise at the surface of the container is obtained by letting the
“spherical” radius of the cylinder be 2 m for a cylinder with a 1-m radius and a 5-m
length. The temperature rise Tat the surface is with respect to the temperature
at a distant location, r = ∞. Using k = 2.1 W/m ◦ C for the volcanic tuff at Yucca
Mountain, we obtain the temperature rise for 30-year old fuel,
T = Pheat /4πkr surface = 12 kW/(4π )(2 m)(2.1 W/m ◦ C) = 230◦ C. (7.53)
This result is consistent with formal calculations. Spent-fuel heating will drop
after 1000 years to allow water to collect and corrode the package, but the package
is designed to deflect the water. The primary radioactive leakage would be the Pu
and Am actinides and not fission fragments, which would have mostly decayed
by that time.
7.7.4 Pu Migration
The calculation of dispersal of SO2 in power plant plumes (Section 6.4) depended
on empirical parameters, such as the weather, air turbulence, and local geography.
Similar corrections would have to be made to estimate radioactive water flow in
underground plumes. Estimates can be obtained using the diffusion equation for
water flow and Darcy’s law, which determines water discharge flux (kg/s m2 ) from
factors of media permeability, fluid density, fluid dynamic viscosity, and ground-
water pressure. Such a calculation can be done in high-porosity media, but it can-
not be done well in low-porosity media, since flow through cracks surpasses flow
through low-porosity pores. It would be misleading to calculate water flow at Yucca
Mountain since its geological stratum is heterogeneous and does not consist of con-
centric layers of homogenous materials. Thus, caveat emptor.
The 1978 American Physical Society panel on radioactive wastes concluded that
“Pu [is] efficiently confined” to regions close to the 1.8-billion year old Oklo, Gabon,
natural reactor. This conclusion was readily accepted because Pu has low solubility
in water and it has a tenacious capacity to cling to mineral surfaces. However, in
the past 15 years there has been some evidence that Pu can adhere to colloids
smaller than a micron in size. In one case a DOE group discovered that Pu had
P1: GFZ
migrated 1.3 km in 30 years from its nuclear weapons test origin. For this case, the
rate of Pu travel was consistent with the local flow of groundwater. The amount
of Pu found was small, some 10−14 mole/l, but it could accumulate over the years.
On the other hand, some argue that the data are misleading, since the migration
might be a result of nuclear weapons explosions that created underground fissures,
increasing Pu transport. Transportation of Pu is only one factor, as there are several
steps to determining human risk, including the size of the population in the region.
Perhaps the issues should be broadened. Perhaps the surface radiation rate (an area
within 15-km radius from Yucca Mountain in 10,000 years) is less relevant than the
number of people affected in Las Vegas and near the Colorado River, as compared
to those affected by alternative disposal approaches.
7.8 Indoor Radon

7.8.1 Radon
One 1984 morning upon arriving at work, an employee at a Pennsylvania nu-
clear power plant triggered the plant’s alarm. It was assumed that he had taken
radioactivity home for the night, but it was soon determined that he went home
without a trace of radioactivity. Surprisingly, what happened was that he carried
radon daughters from home to work. The radon level in his house was 700 times the
EPA-recommended indoor limit of 4 nCi/m3 . The average level inside US buildings
estimated by EPA is 1.3 nCi/m3 , which is about 3.5 times the average outdoor level
of 0.4 nCi/m3 . Six million homes exceed the EPA level, 100,000 of them have lev-
els above 20 nCi/m3 . Radon exposure is a major issue because the average radon
dose of 2 mSv/year (200 mrem/year) is over 50% of the total background rate of
3.6 mSv/year (360 mrem/year). This is partly due to the fact that people spend
86% of their time indoors (with the other 6% in vehicles and 8% outdoors).
The principal health risk from radon arises not from 222 Rn, which does not ad-
here to lungs, but rather its four radioactive daughters (218 Po, 214 Po, 214 Bi, 214 Pb),
which chemically attach to aerosols that are trapped in lungs. Radon concentra-
tion is increased with increased local radon source strength, with reduced air in-
filtration through walls/ceilings that traps radon inside, and with increased air
coupling between radon ground sources and house interiors. (Increased air infil-
tration through walls and ceilings reduces radon, but it also increases infiltration
heat transfer, which is about a quarter of cooling/heating energy.) The radon level
in your home can be measured with a $20 kit. We estimate both energy savings
and adverse health effects from reduced infiltration. One can both save energy and
reduce radon hazards by using air-to-air heat exchangers.
7.8.2 Energy Savings from Reduced Infiltration

Let us estimate the energy savings if the US reduced infiltration from 1.5 to 1.0 air
changes per hour (ach). The infiltration energy loss rate is
d Q/dt = (dm/dt)cT/η, (7.54)
P1: GFZ
7.8. Indoor Radon 191
where mass infiltration rate is dm/dt; specific heat of air is c = 1000 J/kg ◦ C, tem-
perature difference between the inside4 and outside of a house is T = 18.3◦ C −
Toutside , and furnace/duct efficiency is η = 2/3. The rate of air mass infiltration is
dm/dt = NACH Vρ, (7.55)
where the number of air changes/hour is NACH , house interior volume is V, and
air density is ρ = 1.3 kg/m3 . Summing the infiltration energy loss over the year
gives the annual infiltration energy loss (Section 11.3),
Q = NACH VρC(dd/yr × 24 h/day)/η. (7.56)
The number of heating degree days per year (dd/yr) is calculated by summing tem-
perature difference over a year on an hourly basis:
8766
dd/yr = i=1
Ti (1 h)/24 h. (7.57)
Since the average US heating season can be described as having an average tem-
perature of 38◦ F over 6 months, Eq. 7.57 becomes
dd US /yr = Tt = (65◦ F − 38◦ F)(180 days/year) = 4800◦ F day/year
= 2670◦ C day/year. (7.58)
US energy savings are estimated using the assumption of an NACH reduction
from 1.5 to 1 ach in each of the 100 million living units, each with volume V =
325 m3 (area 130 m2 × ceiling height 2.5 m). The US annual energy savings could
be as large as
Qsavings = 108 NACH VρC(dd × 24 h/day)/η
= 108 (1.5 ach − 0.5 ach)(325 m3 )(1.3 kg/m3 )(103 J/kg ◦ C)(2670◦ C dd/yr)
× (24 h/day)(1.5)
= 2.0 × 1018 J = 1.9 × 1015 Btu = 15 quads. (7.59)
This annual energy savings is equivalent to 0.9 million barrels/day of oil or 1.9 tril-
lion cubic feet/year natural gas. Additional savings come in summer from reduced
air conditioning.
7.8.3 Radon Levels in Houses

The slight underpressure inside a house sucks underground radon into houses at
rates between 0.1 to 100 pCi/sec m2 . Equating inward radon flux f over floor area
A to the outward flow gives radon concentration c Rn at one ach (radon decay will
be added later):
f A = (1 pCi/s · m2 )(130 m2 ) = c Rn (V/t) = c Rn (325 m3 /3600 s). (7.60)
4
The US uses 65◦ F (18.3◦ C) as the reference point for calculating degree days. The 68◦ F inside
temperature is reduced by 3◦ F to take into account the inside “free temperature” gained by
the thermal resistance of the house. (Sections 11.3 and 11.5.)
P1: GFZ
Using f = 1 pCi/s m2 , A = 130 m2 , V = 325 m3 , and 1 ach (t = 3600 sec) gives
c Rn = 1.4 nCi/m3 . A house with NACH reduced to 0.35 ach raises c Rn to the suggested
EPA limit of 4 nCi/m3 . Radon’s atomic density n at 1 ach is
nRn = c Rn τ = (1.4 × 10−9 Ci)(3.7 × 1010 /s Ci)(4.7 × 105 s) = 3 × 107 /m3 , (7.61)
using c Rn = 1.4 nCi/m3 and radon mean life τ = T1/2 /ln2 = 3.8 days/0.693 = 4.7 ×
105 s. Radon density is very small, only 10−17 of atmospheric density. For very
low infiltration rates, radon concentration is reduced by 1–10% because of radon
decay in the house. The corrected radon radioactivity density c Rn ’ is determined
by balancing incoming radon rate with the sum of the decay and exhaust rates:
f A = c Rn /τ + c Rn V/t. (7.62)
7.8.4 Cancer Rates

We estimate an increase in radon cancers from reducing infiltration, using results
from the UN Scientific Committee on Radiation (UNSCEAR). Uranium miner data
show that 100 additional lung cancers would be caused every year if 1 million
persons spent all their time in 1 nCi of 222 Rn/m3 . If NACH is reduced from 1.5 to
1 ach, the radon level would be increased from 1 to 1.5 nCi/m3 . The number of
additional cases of lung cancer for 300 million Americans, who are inside 86% of
the time, is about
(1.5 − 1)(nCi/m3 )(10−4 /nCi/m3 )(0.86)(3.0 × 108 persons) = 13,000/year. (7.63)
This is similar to EPA estimates 14,000 radon deaths/year (from 7000 to 30,000)
and National Academy estimates of at 11,000 deaths/year (BEIR VI, from 3000 to
33,000). An alternative approach is to use the 1990 BEIR average radon exposure
estimate of 2.0 mSv/year (55% of the background rate of 3.6 mSv/year). Using this
with the International Commission on Radiological Protection estimate of 20 Sv for
one statistical death (EPA has used 25 Sv), the US fatality rate from radon is
(2.0 mSv/year)(300 million persons)/(20 Sv/death) = 30,000/year. (7.64)
Using these results, EPA recommend a radon limit of 4 nCi/m3 . Radon measure-
ments costing $15 to $50 can determine if radon is a problem. Some 50,000–100,000
houses have radon levels in primary living spaces greater than 20 nCi/m3 . Thirty
years occupancy in such a house can cause a lung cancer rate of about 2–3% among
nonsmokers and 12% among smokers. Remediation can cost $1000–1500 for sub-
slab depressurization by installation of an electric fan and other measures. EPA
estimates that it would cost $19 billion to save some 83,000 lives, or $200,000/life.
Air-to-air heat exchangers can reduce radon and save energy. Heat exchangers
transfer 75% of heat/coolth from exhaust air to incoming air, while radon and
chemicals are released to the outside. To find infiltration leaks, blower doors over-
pressurize houses so house doctors can follow smoke from miner’s smoke sticks to
find and caulk air leaks. The energy and health tradeoff depends on the number of
air exchanges per hour. On the one hand, energy consumption from infiltration is
P1: GFZ
Problems 193
proportional to NACH . On the other hand, cancer rate is proportional to radon con-
centration, which is inversely proportional to NACH . The optimal solution depends
on a comparison of the value of a life as compared to the value of energy saved, a
difficult comparison.
Problems
40
7.1 K universe. How long ago was potassium produced if 39 K and 40 K were
equally produced in early supernovas? Potassium contains a 0.012% (1.2 ×
10−4 ) 40 K, which has a 1.3 billion year half-life while 39 K is stable.
40
7.2 K dose. A typical person receives 0.39 mSv/year (39 mrem/year) from in-
ternal sources, about 50% of this from 40 K. (a) How much potassium is in an
80-kg adult, assuming that 1 MeV per 40 K decay is deposited in the body?
(b) What annual dose does a person absorb while sleeping 0.5 m from a 80-kg
spouse? Assume 50% of the incident 1.5-Mev γ rays, which are emitted in 10%
of the 40 K decays, are absorbed by the body.
7.3 Full-body dose. What is the full-body dose of an 80-kg person ingesting
1 Curie of tritium, which deposits 6 keV per decay. The half-life of tritium
is 12.3 year, but its biological half-life is 10 days.
7.4 MW-days. Show that the fission energy from 1 g of 235 U is 1 MWt -day.
7.5 Nuclear policy. (a) Why did the US choose LWRs? Why did Canada choose
heavy water reactors? (b) Why are heavy water reactors both more and less
dangerous for proliferation of nuclear weapons? (c) What are three main fac-
tors that affect commercialization of breeder reactors? (d) How can proton
accelerators breed plutonium (50 neutrons/GeV) or “burn” actinides? What
are some difficulties?
7.6 Finite, cheap uranium fuel. The United States has about 5 million tons of
inexpensive uranium ore. How many GWe would this sustain if each reactor
were to last 40 years, and 5/7 of the 0.7% 235 U were available from enrichment?
7.7 Single decay. What is the radioactive decay rate in nuclear weapons made
with 5 kg of plutonium, 94% 239 Pu (T1/2 = 24,000 year) and 6% 240 Pu (6,600
year). (b) What would be the rate from today’s weapons in 10,000 year?
7.8 Double decay. One gram of pure 239 U is produced in a reactor. It decays with
a half-life of 24 min to 239 Np, which decays with a half-life of 2.4 days to 239 Pu.
Write and solve the two coupled differential equations that describe the ra-
dioactivity. Sketch a graph of time dependence of the decays and populations
of these isotopes for first 5 days.
7.9 Annual dose of 3.6 mSv. Your 80-kg body gets a dose of 3.6 mSv/year. How
much energy is deposited per year? How many cells does your body contain
if a cell’s average diameter is 20 μ? On average, how much radiation energy
does each cell in receive per year? How many chemical bonds are broken in
each of your cells per year if the bond energy is 5 eV?
7.10 Cosmic ray dose. The average dose from cosmic rays is 0.31 mSv/year
(31 mrem/year). If we assume this dose is mostly from muons with a flux
P1: GFZ
of 100/s·m2 , what is the average energy deposited per muon (decay, kinetic
energy loss, other reactions) in an 80-kg person with a 0.1 m2 area? How does
this compare to 106-MeV muon decay energy and 3-GeV kinetic energy?
7.11 Mass of nuclear waste. (a) How much nuclear waste does a 1-GWe plant
produce each year if it is 32% efficient with an 85% load factor? What is the
volume of waste, with and without reprocessing? (b) What is the waste activity
after 5 years storage if the average lifetime of all fission fragments is 20 years
and each fragment decays 2–3 times?
7.12 240
Pu/239 Pu/238 U. Uranium is placed in a reactor with neutron flux 1014 /cm2 ·s.
What are the 239 Pu/238 U and 240 Pu/239 Pu ratios after 2 months with thermal
σ (238 U) = 2.7 barns and σ (239 Pu) = 271 barns?
7.13 Passively safe reactors. What amount of carbon in a 1-GWe , high temperature
gas reactor would keep its temperature below 1700◦ C after a LOCA?
7.14 Fault-tree analysis. (a) Design a fault tree with the following features: five
sensors to detect coolant loss, each with successful probability of 99%; two
sets of electrical connections to open the emergency cooling water at 99%
each; two valves at 98% each; the availability of line power at 99%; two back-
up generators at 98% each; the presence of water 99.9%. What is the probability
of failure? What common-mode failures from an earthquake does this ignore?
7.15 Breeder doubling time. A breeder reactor with a plutonium core creates about
2.5 neutrons/fission with 1 neutron to maintain the power, 1 to convert 238 U
to 239 Pu, and 0.5 going to losses. (a) How long would it take a 3-GWt reactor
to produce 3 tons of plutonium, recalling that 1 MWt -day = 1 g 235 U. (b) What
is the Pu gain each year? (c) How long would it take to double the 4-ton Pu
core?
7.16 Liquid metal cooling. To avoid geometrical neutron-spreading that raises the
cost of Pu, the core of a breeder must be compact with 20% fissile content. To
remove 3 GWt of heat, a large flow of liquid sodium cools the core, a process
that avoids high-pressure water-cooling. (a) What is the outlet temperature of
sodium if 10 m3 /s enters at 620◦ C? (b) What is sodium velocity with pipe areas
0.1 m2 to 1 m2 ? (The specific heat of Na at 620◦ C is 1.3 J/g-◦ C with density
800 kg/m3 .)
7.17 Coal vs. nuclear. Discuss and quantify the health, safety, and environmental
parameters for a comparison of power production from 100 GWe of coal versus
100 GWe of nuclear.
7.18 Wedge plume with threshold. (a) Redo the APS calculation for reactor acci-
dents, but with a very arbitrary choice of 0.01 Sv (1 rem) for a threshold dose
of radiation. (Below 0.01 Sv radiation is arbitrarily assumed not to be harm-
ful.) (b) The APS calculations assumed a person lived continually in a region
with 137 Cs on the ground. How did the SLBM accident calculation differ from
this?
7.19 Radon guideline. The EPA recommends radon be kept below c Rn = 4 nCi/m3 .
(a) What is the decay/m3 for 222 Rn? (b) How many 222 Rn nuclei are in a cubic
meter? (c) A daughter of 222 Rn has a half-life of 30 min. How does this affect
your answer to the previous question? (d) What is the concentration of 222 Rn?
P1: GFZ
Bibliography 195
7.20 Radon and energy. Is it economically viable to reduce infiltration in Chicago

with 6000◦ F-day/year from 1.5 ach to 1 ach if a human life is valued at $5 mil-
lion? Assume infiltration energy saved is 10 MBtu/year and natural gas costs
$5/MBtu and electricity costs $14/MBtu.
7.21 Correction for radon decay. Redo the text calculation of radon density for
0.1 and 1 ach, but include radon decay in the house.
Bibliography
Ahearne, J., et al. (1997). Nuclear waste issues, Phys. Today 50, 22–66.
Alverez, R., et al. (2003). Reducing the hazards from stored spent power reactor fuel in the
United States, Sci. Global Secur. 11, 1–51.
American Physical Society (1975). Light Water Reactor Safety, Rev. Mod. Phys. 47, S1–S124.
———(1977). Nuclear Fuel Cycle and Waste Management, Rev. Mod. Phys. 50, S1–S186.
———(1985). Radionuclide Release from Severe Accidents at Nuclear Power Plants, Rev.
Mod. Phys. 57, S1–S154.
Barkenbus, J. and C. Forsberg (1995). Internationalizing nuclear safety, Ann. Rev. Energy
Environ. 20, 179–212.
Bodansky, D. (2004). Nuclear Energy, American Institute of Physics Press, New York.
Boeker, E. and R. Grondelle (1995). Environmental Physics, Wiley, New York.
Bupp, I. and J. Derian (1978). Light Water, Basic Books, New York.
Cochran, T., et al. (1993). Radioactive contamination at Chelyabinsk-65, Russia, Ann. Rev.
Energy Environ. 18, 507–528.
Craig, P. (1999). High-level nuclear waste: The status of Yucca Mountain, Ann. Rev. Energy
Environ. 24, 461–486.
Eisenbud, M. and R. Gesell (1997). Environmental Radioactivity, Academic Press, New York.
Fetter, S. and F. von Hippel (1990). The hazard from plutonium dispersal by nuclear-warhead
accidents, Sci. Global Secur. 2, 21–41.
Flynn, F. and P. Slovic (1995). Yucca mountain: Prospects for America’s high-level nuclear
waste program, Ann. Rev. Energy Environ. 20, 83–118.
Forsberg, C. and A. Weinberg (1990). Advanced reactors, passive safety and acceptance of
nuclear energy, Ann. Rev. Energy Environ. 15, 133–152.
Golay, M. (1993). Advanced fission power reactors, Ann. Rev. Particle Sci. 43, 297–332.
House of Representatives (1990). Nuclear Weapons Safety (Drell), HASC-15, House of Repre-
sentatives, Washington, DC.
Lamarsh, J. (1977). Introduction to Nuclear Engineering, Addison Wesley, Reading MA.
Levi, M. and Kelly, H. (March/April 2002). Dirty bombs: response to a threat, Public Interest
Report, Federation of American Scientists
Mabuchi, K. (1998). Effects of Ionizing Radiation: Atomic Bomb Survivors, L. Peterson and
S. Abrahamson (Eds.), Joseph Henry Press, Washington, DC.
MacFarlane, A. (2001). Interim storage of spent fuel in the United States, Ann. Rev. Energy
Environ. 26, 201–236.
Marcus, G. and A. Levin (2002). New designs for the nuclear renaissance, Phys. Today 55(4),
54–60.
Mettler, F. and A. Upton (eds.) (1995). Medical Effects of Ionizing Radiation, Saunders, Philadel-
phia, PA.
P1: GFZ
National Research Council (1990). Health Effects of Exposure to Low Levels of Ionizing Radiation,
National Academy Press, Washington, DC.
———(1992). Nuclear Power: Technical and Institutional Options for the Future, National
———(1999). Health Effects of Exposure to Radon, National Academy Press, Washington, DC.
———(2000). Long-Term Institutional Management of US Department of Energy Legacy Waste
Sites, National Academy Press, Washington, DC.
Nero, A. (1979). Guidebook to Nuclear Reactors, University of California Press, Berkeley, CA.
Nuclear Regulatory Commission (1975). Reactor Safety Study (Rasmussen), Nuclear Regula-
tory Commission, Washington, DC.
Office of Technology Assessment (1991). Complex Cleanup: The Environmental Legacy of Nuclear
Weapons Production, OTA, Washington, DC.
Peterson, L. and S. Abrahamson (eds.) (1998). Effects of Ionizing Radiation: Atomic Bomb Sur-
vivors and Their Children, Joseph Henry Press, Washington, DC.
Rasmussen, N. (1981). The application of probabilistic risk assessment techniques to energy
technologies, Ann. Rev. Energy Environ. 6, 123–138.
Roberts, R., R. Shaw and K. Stahlkopf (1985). Decommissioning of commercial nuclear power
plants, Ann. Rev. Energy Environ. 10, 251–284.
Schillaci, M. (1995). Radiation and Risk (UNSCEAR94 data), Los Alamos Sci. 23, 104–105.
Wilson, R. (1999). Effects of ionizing radiation at low doses, Am. J. Phys. 67, 322–327.
P1: GFZ
8
Climate Change
8.1 Introduction
Weather can change many times a day. Climate—the sum of weather—changes
slowly, over decades and centuries, but it can change abruptly with large volcanic
eruptions, instabilities in ocean currents, or meteorite crashes. The dramatic 1815
Tambora eruption spewed 100 km3 of ash, causing “a year without a summer”
to cool Earth by 4◦ C. Cooling from volcanic and anthropogenic aerosols must be
factored into climate predictions. Without industrialization, Earth’s temperature is
raised 33◦ C by greenhouse gases [water vapor (21◦ C), CO2 (7◦ C), other gases (5◦ C)].
The Intergovernmental Panel on Climate Change (IPCC) projected in 2002 an addi-
tional 2.5◦ C rise (1.4–5.8◦ C) for a doubled CO2 (560 ppm). Positive feedback magnifies
warming. For example, increased CO2 warms Earth, which increases atmospheric
moisture that further raises Earth’s temperature. Positive feedback also comes from
reduced snow and ice, which reduce solar reflection. Increased cloudiness has a
negative feedback since increased numbers of high-altitude clouds increase solar re-
flection. But more low-altitude clouds also produce positive feedback by trapping
infrared radiation. The height of the clouds is critical in this comparison.
Few scientists debate the fact that Earth is getting warmer. Seventeen of the eigh-
teen warmest years in the 20th century occurred after 1980. The IPCC determined
that Earth’s surface temperature rose 0.6◦ C (+/− 0.2◦ C) during the 20th century.
Half this increase occurred since the mid-1970s to reach the highest temperatures in
a millennium. Reanalysis of satellite tropospheric microwave emission data is con-
sistent with this result, giving a rise of 0.1◦ C per decade between 1979–2001. Mea-
surements (2003) show that the tropopause, the region between the stratosphere
and the troposphere, rose 200 m in the past two decades. The authors attribute
more than 80% of this change to human activity. The IPCC further estimates a rise
of 1.4–5.8◦ C by 2100 if carbon emission trends are not significantly reduced. Such
a temperature rise would be significant, similar in magnitude to the 5–10◦ C drop
of glacial periods. See Fig. 8.1.
The oceans rose 15 cm in the past century as glaciers melted and warmed oceans
expanded. Scientists expect that high latitudes would show the first signs of warm-
ing. This, in fact, happened as Alaska’s temperature rose by 2◦ C over the 1980s and
197
P1: GFZ
198 8. Climate Change
Figure 8.1. 1000 years of global CO2 and temperature change. Northern Hemisphere surface
temperatures, CO2 concentrations, and carbon emissions are strongly correlated. Tempera-
ture records were obtained from historical records, tree rings, and corals. CO2 concentrations
are obtained from air bubbles in layered ice cores drilled in Antarctica and from atmospheric
measurements since 1957. Carbon emissions are reconstructed from fossil fuel combustion
and land-clearing data since 1750(US Global Climate Change Research Program, 2000).
P1: GFZ
8.1. Introduction 199
1990s, and the North Polar Cap thinned by 40% (from 3.1 m to 1.8 m) in 30 years,
along with area reduction of 3% per decade. In the 40–60◦ N latitudes, the first
day for lake-freezing shifted an average of 6 days per century to later in winter,
while the ice breakup day moved earlier at a rate of 6.5 days per century. The South
Polar region showed similar trends as the Wordie Ice Shelf decreased from 2000
to 700 km2 between 1966 and 1989. As tundra warms, it releases methane, CH4 ,
which is 15 times per molecule more effective at trapping infrared compared to
CO2 . Methane provides a positive feedback by further increasing temperatures by
releasing methane from warmed tundra. But warmed tundra also has a negative
feedback because it traps CO2 by increasing plant growth. It is expected that the
higher latitudes would show the first signs of global warming, but southern lakes,
such as Lake Tanganyika, have reduced vertical mixing, increasing warming rates.
It is projected that 55% of future additional radiation forcing (W/m2 ) will be caused
by CO2 ; 15% will be caused by CH4 and 30% by other gases. There is little doubt
that Earth is warming and that carbon dioxide is connected to this temperature
rise. But not all parties accept IPCC estimates for the amount of temperature rise
from increased carbon dioxide concentrations.
Earth experiences major climate changes driven by several types of mechanical
cycles, called Milankovitch cycles: a cycle of 26,000 years is due to spin precession;
a 40,000-year cycle from the 3◦ variation in spin axis angle; and cycles of 100,000
and 400,000 years are due to orbit eccentricity. The Earth will cool in some 10,000
years, as a result of Milankovitch cycles but here we deal only with near-term
changes caused by the burning of carbon fuels. A connection between CO2 and
increased temperatures seems likely, but to what extent is CO2 responsible? Data
taken from entrapped air in ice samples between the glacier’s surface and depths
of 2 km in Greenland and Antarctica show that CO2 and methane concentrations
are strongly correlated with changes in isotopic ratios (2 H/1 H and 18 O/16 O) over
the past 150,000 years. As CO2 rose from 190 to 300 ppm and CH4 rose from 300
to 700 ppb, the amount of the heavier H and O isotopes increased, making for a
temperature rise by 10◦ C. Does correlation between CO2 and temperature constitute
a sufficient proof to claim that raised temperatures are caused by increased CO2 ? Or
does this correlation merely show a coincidence to another cause? The solar cycle
might bring about climate change to some extent, raising both temperature and
CO2 , but this is not a strong theory. However, there is a consensus among scientists
who work in climate change research as represented in the position taken by the
American Geophysical Union in its 1999 report “Climate Change and Greenhouse
Gases;”
“. . . There is no known geologic precedent for the transfer of carbon from the Earth’s crust to
atmospheric carbon dioxide in quantities comparable to the burning of fossil fuels without
simultaneous changes in other parts of the carbon cycle and climate system. This close
coupling between atmospheric carbon dioxide and climate suggests that a change in one
would in all likelihood be accompanied by a change in the other.”
Atmospheric CO2 is now at its highest in the past 400,000 years and it is projected
to rise in the 21st century from its preindustrial value of 280 ppm (now at 370 ppm)
P1: GFZ
to 600–800 ppm (highest in 40 million years). These estimates (2001, EIA) are based
on annual carbon use rising from today’s 6.5 gigatonnes (1 Gton = 1 Gt = 109 tons)
to 7.7 Gt in 2010, 10.4 Gt in 2025 and 20 Gt in 2100. The developing countries are
projected to fill 58% of this growth. Temperature changes from other factors should
also be considered, such as solar variations, Earth spin-orbit oscillations, scattering
by aerosols and uncertain cloud contributions.
Does theory lead experiment or does experiment lead theory? General circula-
tion model (GCM) calculations are fairly successful at extrapolating past climate
changes. GCM calculations do not include all the effects from oceans, clouds, and
biosphere, but they do allow theory to lead experiment by giving estimates based
on CO2 levels. Measurements can be ambiguous since results can be attributed
to several causes. Our goal is not to answer the big question of whether CO2
will cause severe climate change, but only to estimate some aspects of climate
change.
Control of CO2 emitted by cars and industry is a hotly debated issue. The United
States consumes carbon at the rate of 5.5 ton/person year (1.56 Gt/284 million
in 2001, EIA), twice that of Europe and Japan, and much higher than the global
rate of 1 ton/person year (6.5 Gt/6.3 billion population). Carbon consumption in
the United States during 2001 by sectors was transportation (0.51 Gt), industry
(0.45 Gt), residences (0.32 Gt), and commerce (0.28 Gt). US carbon use is projected
to grow from 1.56 Gt in 2001 to 2.2 Gt in 2025 with a 1.5%/year growth rate. Global
carbon use is projected to grow from 6.5 Gt in 2001 to 10.36 Gt in 2025 with a
1.9%/year growth rate. Most of the growth in carbon consumption will be in the
developing nations and not in the industrialized states.1 In 2001, President George
W. Bush withdrew the United States from the Kyoto Protocol for the stated reason
that it imposed constraints on the United States to which China and India would
not be subjected. Today the nations in the Organization for Economic Cooperation
and Development (OECD) produce 51% of CO2 while the Kyoto nonsignatories
produce 35%. However, these percentages will change as China and India expand
their carbon consumption (Section 8.9). The Kyoto Protocol entered into force on
February 16, 2005, after Russian ratification. The 120 nations that ratified emitted
61% of greenhouse gases by signatories, over the 55% threshold for ratification
with the United States emitting 37%. The carbon-trading allotments initially were
selling for $10 a ton. See Figs. 8.2 and 8.3.
Some US states are entering climate change policy at their level of government.
The California Air Resources Board required automobiles to lower carbon emis-
sions by 30%, to be phased in over 2009 and 2016. The Board claimed the extra cost
would be $1000 per vehicle, but it would save $2500 in fuel. New York and other
states claim that they will follow suit.
1
The regional composition is as follows: (International Energy Outlook, EIA, 2003, www.eia.
doe.gov)
2001: industrialized (3.2 Gton), developing (2.3 Gton), former Warsaw Pact (1.0 Gton)
2025: industrialized (4.3 Gton), developing (4.5 Gton), former Warsaw Pact (1.5 Gton).
P1: GFZ
8.2. CO2 Projections 201
Figure 8.2. World carbon diox-

ide emissions by fuel type. Data
are from 1970–2000 in 109 tons
carbon per year (Gton-C/year).
[Energy Information Adminis-
tration, 2001]
Figure 8.3. The global carbon cycle (Office of Technology Assessment, 1991).
8.2 CO2 Projections

The world burned 6.5 Gt of carbon in 2001, with the United States accounting for
1.56 Gt, divided between electricity (0.65 Gt), transportation (0.51 Gt), and industry
(0.36 Gt). In addressing ways to reduce this, scientists talk about enhanced end-use
efficiency and population growth, but they avoid talk of life style, while other groups
may condemn lifestyle but avoid talk of population growth. As a nation becomes
wealthier, their its have fewer children, but each citizen consumes more energy and
goods. Symbolically, carbon production rate in atmospheric CO2 is
C = P E cap /ηC + NC , (8.1)
where C is annual rate of carbon emissions, P is global population, E cap is annual
P1: GFZ
per capita energy consumption, ηC is efficiency of obtaining useful energy from

consumed carbon, and NC is net carbon production from all natural causes. The
product of P and E cap in 2000 was 400 quads (4 × 1017 Btu) or 190 Mbbl/day of oil
equivalent, of which 85% was from fossil fuels. Global per capita fossil energy use
for 6.3 billion inhabitants is
E cap = 0.85(190 Mbbl/day)(365 day/year)/(6.3 billion persons) = 9 barrels/year.
(8.2)
For fossil fuels at 75% carbon by weight, the global annual per capital consump-
tion of carbon is
(0.75)(9 bbl/year)(42 gal/bbl)(3.5 kg/gal) = 1 ton C/person-year. (8.3)
US energy consumption is 100 quads/year, with 85% from fossil energy, giving
a rate 5.5 times higher than the global rate; that is,
E cap = 0.85(47 Mbbl/day)(365 day/year)/(300 million persons) = 50 barrels/year,
(8.4)
or 5.5 ton C/person year. The value of ηC varies between zero (a house fire with
no useful energy) to medium values for electricity derived from coal (better values
with natural gas using combined cycle gas turbines), to very large values from
fission, solar and fusion. Energy has both a supply-production side and a demand-
use side. Improved end-use efficiency (Chapter 14) raises ηC . For example, new
refrigerators use one-third the energy of previous models, effectively increasing ηC
by a factor of 3. The world obtained 85% of its 2000 energy from burning 6.5 Gt of
carbon giving an average fuel-to-carbon ratio of
fuel/carbon = 0.85(190 Mbbl/day)(365 day/year)/(6.5 Gt C/year) = 9 bbl/ton C.
(8.5)
The ratio of fuel mass burned to carbon mass released is about
Mfuel /Mcarbon = (9 bbl)(42 gal/bbl)(3.5 kg/gal)/(1000 kg C) = 1.3 kg/1 kg C. (8.6)
This is similar to octane’s (C8 H18 ) fuel to carbon mass ratio (114/96 = 1.2), natural
gas’s ratio (CH4 , 16/12 = 1.3), and coal’s ratio (1.3, but varies greatly). However,
these numbers are somewhat misleading because hydrogen in natural gas and
petroleum give extra energy that does not produce carbon dioxide. The differential
carbon rate (Eq. 8.1) with speculative estimates from recent global data is
Ċ/C = Ṗ/P + Ė cap /E cap − η̇C /ηC + ṄC /NC (8.7)
Ċ/C = 1.2%/year + 0.8%/year − 0.5%year + 0 = 1.5%/year. (8.8)
For simplicity we ignore changes in natural effects with dNC /dt = 0. (This ap-
pears to be almost true, as forest growth now about matches forest destruction.)
The fractional increase in global population, (dP/dt)/P, in 2002 is 1.2%/year. The
fractional growth of global energy use, (dE/dt)/E, was 2%/year in 2001 with a per
capita growth (dEcap /dt)/E cap of 0.8%/year. Energy efficiency growth (dηC /dt)/ηC
P1: GFZ
8.2. CO2 Projections 203
Table 8.1. Carbon released in 1999 from petroleum,

natural gas, and coal (EIA, 2002)
Petroleum: (152 quad)(0.018 Gton/quad) 2.7 Gton

Natural Gas (87 quad)(0.015 Gton/quad) 1.3 Gton
Coal: (85 quad)(0.025 Gton/quad) 2.1 Gton
Total 6.1 Gton
was 1%/year in the United States in the 1980s, but it is less now at 0.5%/year.
Combining these factors gives a carbon growth rate of about 1.5%/year.2
8.2.1 Projections
To understand CO2 sources and sinks we estimate CO2 emissions into the atmo-
sphere and compare to measured growth rates. Global energy consumption in 1999
was 382 quads (403 exajoules3 ) with fossil fuels supplying 85% (324 quads) with
petroleum (152 quads), natural gas (87 quads), and coal (85 quads). On an energy
basis, natural gas produces the least amount of CO2 , petroleum produces a median
amount, and coal produces the most CO2 , as seen in Table 8.1.
The 1999 EIA value of 6.1 Gt of carbon was much less than the amount trans-
ferred between the atmosphere and terrestrial ecosystems (120 Gt/year each way)
or between the atmosphere and the oceans (90 Gt/year each way). The atmosphere
has a sink rate of about 3.5 Gt/year and land has sink rate of about 1.5 Gt/year. The
oceans sink rate of about 2 Gt/year sequestered some 100 Gt of anthropogenic car-
bon over the industrialized years. The carbon released from deforestation is about
1 Gt/year, raising the total carbon released to the atmosphere to about 7.1 Gt/year
in 1999. The number of CO2 molecules released per year is
NCO2 = (7.1 × 1015 g/year)(6.02 × 1023/mole)/(12 g/mole) = 3.6 × 1038 CO2 /year.
(8.9)
To convert this into parts per million in the atmosphere, we need Earth’s atmo-
spheric mass to obtain the number of molecules in the atmosphere. The mass of the
atmosphere is the total weight of the atmosphere (area of Earth 4π R2 × atmospheric
pressure P = 105 Pa) divided by g:
Matmos = P A/g = (105 Pa)(4π )(6.4 × 106 m)2 /(9.8 m/s2 ) = 5.3 × 1018 kg. (8.10)
This gives the total number of O2 and N2 molecules in the atmosphere,
Natmos = (5.3 × 1021 g)(6.02 × 1023 /mole)/(29 g air/mole) = 1.1 × 1044 molecules.
(8.11)
2
I confess to raising (dC/dt)/C with three children, (dP/dt)/P, and raising (dEcap /dt)/E cap
by visiting Parisian grandchildren.
3
1 quad = 1015 Btu = 1.055 × 1018 J = 1.055 exajoules = 1.05 EJ. The global 382 quads =
403 EJ.
P1: GFZ
This gives an increase in the CO2 molecular concentration,
NCO2 /Natmos = (3.6 × 1038 CO2 /year)/(1.1 × 1044 air) = 3.3 ppm/year. (8.12)
This is more than twice the atmospheric CO2 rise of 1.4 ppm/year (325.3 ppm in
1970 to 354 ppm in 1990 to 370 ppm in 2000). Thus, about half of the CO2 remains
in the atmosphere, the other half goes into oceans and forest growth. A path for
atmospheric CO2 travel to Earth’s surface is acid rain. The rain from pristine air
at 370 ppm has a 5.5 pH as CO2 becomes H2 CO2 . The sink for the “lost” carbon
is uncertain, but some evidence points to growth in northern hemisphere forests.
The IPCC gave evidence to this uncertainty when they reported a 46% atmospheric
sink (3.3 Gt of 7.1 Gt emissions), a figure that did not account for 1.3 Gt. The lost
CO2 flux is undoubtedly captured by oceans and forests.
8.2.2 CO2 Before Industrialization

Preindustrial CO2 level from analysis of ice core measurements was determined
to be 280 ppm for the year 1800. We now estimate this level beginning with the
1959 level of 315.8 ppm. Before the oil embargo of 1973–74, US energy use was
growing at 4.4%/year, similar to the global rate. The net CO2 molecule lifetime in
the atmosphere is more than a century, which allows us to integrate backward in
time without too much error. Using a rate of 0.9 ppm/year in 1959 and a global
carbon rate growth rate of λ = 3%/year, the increase in CO2 concentration between
1800 and 1959 was
0
c CO2 = (0.9)e λt dt = 0.9(e 0 − e −∞ )/λ = 0.9/0.03 = 30 ppm. (8.13)
−∞
Subtracting this from the 1959 value of 315.8 ppm gives a preindustrial CO2 level
of 285 ppm, close to the accepted value of 280 ppm.
8.2.3 CO2 in the 21st Century

The CO2 level in the middle of the 21st century may be obtained by projecting
60 years growth onto the 1990 level of 354 ppm. The Energy information Agency
estimates that a business-as-usual approach will give 2%/year growth in fossil
fuels, or a CO2 concentration in 2050 of
60
c CO2 = (1.4 ppm/year)e0.02t dt + 354 ppm = 162 ppm + 354 ppm = 516 ppm.
0
(8.14)
These results are similar to the IPCC estimates of 500 to 950 ppm for the year 2100.
The conventional wisdom is that CO2 will double in this century to over 560 ppm,
probably after 2050. See Table 8.2 for other estimates.
P1: GFZ
8.3. Upper-Atmospheric and Surface Temperatures 205
Table 8.2. Estimates of CO2 as a function

of time and growth rate
Year 0%/yr 1%/yr 2%/yr
2050 438 ppm 469 ppm 516 ppm
2100 508 ppm 635 ppm 916 ppm
8.2.4 Alternative Technologies

CO2 production can be reduced with adoption of solar, nuclear, wind, and geother-
mal energy sources. More important, enhanced end-use efficiency outpaced alter-
native energy since it reduced energy use in US cars, buildings, and appliances
by 50%. A further reduction of 50% could be obtained on a cost-effective basis.
The production side has improved with the advent of combined-cycle gas turbines
operating at 60% efficiency. Lastly, CO2 could be sequestered in the deep ocean, in
geological cavities, or vegetation, but the economic aspects of such technologies
have to be examined (Section 8.8). Governments can encourage choices, but its
ability to transform the global economy is limited.
If noncarbon sources were to become economically feasible, market penetration
of these measures would take place over decades, and such penetration may not be
decisive. Many of the enhanced end-use efficiency options are already cost effective,
but there are social barriers slowing their adoption. Assume naively that the fraction
of energy from noncarbon technologies increases linearly as f = t/T, where T is
the time period for total market penetration. The concentration of CO2 after T years
of market penetration after 2000 to an energy era without CO2 production is
T
c CO2 = (1.4 ppm/year)(eλt )(1 − t/T)dt + 370 ppm
0
= (1.4 ppm/Tλ2 )(eλT − 1 − λT) + 370 ppm. (8.15)
Using an optimistic time period for market penetration from 2000 to 2060 (T =
60 year) with a variable energy growth λ from zero to 2%/year, we obtain CO2
levels of 396 ppm (λ = 0%), 406 ppm (1%) and 419 ppm (2%) levels for the year
2060. These levels are considerably less than projections of levels obtained when
alternative and end-use technologies are absent
8.3 Upper-Atmospheric and Surface Temperatures

8.3.1 Upper-Atmospheric Temperature Ta
Earth’s temperature is determined from a heat balance between absorbed energy
from solar flux so = 1367 W/m2 (see Chapter 12) and infrared emission to space.
The solar power intercepted by the area of Earth’s disk (π RE 2 so ) is distributed over
the entire spherical area (4π RE 2 ), giving an average solar flux of so /4 = 1367/4 =
342 W/m2 . Of this, 70% is absorbed by the Earth, and 30% is reflected (the Earth’s
P1: GFZ
Figure 8.4. Global mean energy flows between the surface and atmosphere (Trenbeth, 1996).
albedo a = 0.3), giving an average flux absorbed by the surface and atmosphere,
sabsorbed = (1 − a )(so /4) = (1 − 0.3)(1367/4) = 239 W/m2 . (8.16)
Absorption by clouds and atmosphere reduces solar flux at the surface to an av-
erage of about 200 W/m2 . The energy absorbed by Earth’s surface is sent upward
by infrared, evaporation, and air currents, which is captured by the atmosphere
or passes directly to space. In our first model, we assume that all the absorbed en-
ergy is reradiated to space as IR from a thin surface at the top of the atmosphere
(Fig. 8.4).
The power balance at the top of the Earth’s upper atmosphere is
Pin = (1 − a )(π RE 2 so ) = Pout = εσ Ta 4 (4π RE 2 ), (8.17)
where the temperature of the top of the atmosphere Ta is in Kelvin, σ is the Stefan–
Boltzmann constant, 5.67 × 10−8 W/m2 K4 , and ε is emissivity. As a first step, we
assume the Earth is a blackbody with ε = 1 giving an upper atmosphere tempera-
ture,
Ta = [(1 − a )so /4σ ]1/4 = [239/σ ]1/4 = 255 K = −18◦ C = 0◦ F. (8.18)
The 255 K is the temperature in the middle of the troposphere, 5 km above the
surface (and at 50 km). This is 32 K colder than the average surface temperature of
287 K (14.0◦ C with 1997 averages of 14.6◦ C in the northern hemisphere and 13.4◦ C
in the southern hemisphere).
P1: GFZ
As a comparison we calculate Ta-V for Venus, which has a higher solar flux since
it is 40% closer to the sun as compared to Earth:
so-V = so (dE /dV )2 = (1367 W/m2 )(1.50 × 108 km/1.08 × 108 km)2 = 2610 W/m2 .
(8.19)
However, Venus’s higher albedo of 0.76 reflects a greater fraction of sunlight,
greatly reducing the average absorbed flux to
(1 − a )s0-V /4 = (1 − 0.76)(2610 W/m2 )/4 = 157 W/m2 , (8.20)
which is smaller than Earth’s 239 W/m2 . The upper atmospheric temperature of
hot Venus,
Ta-V = [157 W-m−2 /σ ]1/4 = 229 K, (8.21)
is 26 K colder than Earth’s 255 K. However, Venus’s plentiful CO2 traps IR, giving
it a surface temperature of 750 K, three times Earth’s surface temperature of 287 K.
8.3.2 Surface Temperature Ts

Our zero-dimensional box model did not take into account the following variable
factors:
r reflection, absorption, and emission by air, aerosols, clouds, and surface;
r convection of sensible and latent (evaporation) heat;
r coupling to oceans and ice;
r variations in three dimensions (r , θ, φ);
r variable solar flux (θ , φ, t).
Initially we estimate the surface temperature without considering Ta . We assume

that all the solar flux that is not reflected is transmitted through the air and totally
absorbed by the Earth’s surface f absorbed = (1 − a )so /4. The warmed surface ra-
diates as a blackbody a fraction f IR of the absorbed energy in the infrared, with
the remainder rising in air currents or evaporated moisture. This IR flux is totally
absorbed by the atmosphere, which contains greenhouse gases. The atmosphere
radiates 50% of the IR absorbed flux to space and 50% to Earth, giving an IR flux
downward of ( f IR /2) f absorbed . Again, a fraction f IR of this radiates in the IR to the
atmosphere and 50% of this radiates downward and is absorbed by the surface,
( f IR /2)( f IR /2) f absorbed . This process gives an infinite sum in the energy balance:
f absorbed + ( f IR /2) f absorbed + ( f IR /2)2 f absorbed + ( f IR /2)3 f absorbed + · · ·
+ ( f IR /2)n f absorbed = σ T s 4 . (8.22)
This factors into
f absorbed [1 + ( f IR /2) + ( f IR /2)2 + ( f IR /2)3 + · · · + ( f IR /2)n ] = σ T s 4 . (8.23)
Using the identity 1/(1 − ε) = 1 + ε + ε + ε + · · · + ε , this becomes
2 3 n
f absorbed /[(1 − ( f IR /2)] = σ Ts 4 . (8.24)

P1: GFZ
We obtain the actual Earth surface temperature of Ts = 287 K with f IR = 0.76. For
the extreme case of no IR emission from the surface ( f IR = 0), we obtain Ts = 255 K,
the temperature of the upper atmosphere. For the other extreme case of 100% IR
emission from the surface ( f IR = 1), we obtain Ts = 303 K, consistent with the next
calculation.
8.3.3 Ta and Ts Together

We increase the sophistication of the model by coupling the temperatures for Earth’s
surface and atmosphere. We assume that all sunlight is transmitted through the air
and absorbed by the Earth’s surface. We vary the one free parameter, the emissivity
of the atmosphere ε a , but retain the surface of Earth as a blackbody with ε E = 1.
The first equation balances heat flow in the single layer of air. The left side doubles
the infrared flux emitted by Earth’s atmosphere, since IR goes both up into space
and down to Earth’s surface. This is balanced with IR flux emitted from Earth’s
blackbody surface and absorbed by the gray body atmosphere. This energy balance
and the next one link atmospheric temperature Ta with surface temperature Ts :
2εa σ T a 4 = εa σ T s 4 . (8.25)
The second equation is an energy balance at Earth’s surface. The left side is the sum
of solar energy absorbed at the surface and the absorbed downward flow of IR
from the atmosphere, which is balanced with upward IR flux from the surface,
(1 − a )so /4 + εa σ T a 4 = σ T s 4 . (8.26)
Solving these equations gives
Ts = 21/4 Ta (8.27)
σ T 4s = (1 − a )so /4(1 − εa /2). (8.28)
If the air layer is a blackbody (εa = 1, considerable CO2 ), the atmosphere is Ta =
255 K (as before) and the surface is Ts = 303 K (16 K warmer than actual value of
287 K). If εa = 1/2 (from less CO2 ), the atmosphere is too cold at Ta = 230 K and the
surface is also too cold at Ts = 274 K. By adjusting ε a to 0.76, we obtain the “correct”
surface temperature, Ts = 287 K.
8.3.4 MultiLayer Atmosphere

Following others, we divide the planetary atmosphere into n zones layered verti-
cally. The large temperature difference between the top of the atmosphere and the
surface is troublesome since air temperature in one layer is not constant due to its
large temperature gradient. By using several layers, the temperature gradient in
each layer is reduced, smoothing the temperature profile to become more contin-
uous. The thickness of a layer is such that almost all incident IR on a layer is just
absorbed in that layer, which then radiates it upward and downward. Planets with
small amounts of CO2 and H2 O have less than one zone, while Venus has many
P1: GFZ
zones. To simplify the mathematics we make some drastic assumptions about the
steady-state conditions on a planet:
r All reflection of incident solar flux so takes place at the top of a thin layer (the
clouds of layer zero) of the atmosphere, with the fraction reflected given as the
albedo a . This is a fairly good assumption for Venus, with its copious clouds, but
not very good for Earth.
r All visible solar fux that passes through the top layer is absorbed at the planet’s
surface.
r All atmospheric layers below the top (zeroth) layer totally transmit incident visible
solar flux to the surface.
r All atmospheric layers are just thick enough to totally absorb all incident IR. A quan-
tum of IR will be absorbed just once in a layer to give it one radiation thickness. The
layer absorbs IR coming from the layers just above and just below.
r All atmospheric layers emit IR as a blackbody at the temperature of the layer in
two directions, both up and down.
The temperature at the top of the atmosphere, the zeroth layer of n layers (0 to n − 1),
is denoted as T0 . The energy balance is the same as in Eq. 8.17:
(1 − a )so /4 = σ T o 4 . (8.29)
The left side is the solar flux absorbed at the top of the atmosphere and the right
side is the total IR emission at the temperature T0 . The energy balance in the zeroth
zone is
2σ T 0 4 = σ T 1 4 , (8.30)
where the left side is the upward and downward IR radiation from the zone, and
right side is the IR entering from the n = 1 zone below. Similarly, the heat balances
in the second and third zones are
2σ T 1 4 = σ T 0 4 + σ T 2 4 (8.31)
2σ T 2 4 = σ T 1 4 + σ T 3 4 . (8.32)
The last zone (n − 1) has energy coming from above at Tn−2 and from the surface
below at temperature Ts ,
2σ T n−1 4 = σ T n−2 4 + σ T s 4 . (8.33)
It follows that
T0 = [(1 − a )so /4σ ]1/4 (8.34)
T1 = 21/4 T0 , T2 = 31/4 T0 , Ts = (n + 1)1/4 T0 . (8.35)
Earth’s so = 1367 W/m2 and a = 0.3 gives T0 = 255 K, T1 = 303 K, T3 = 361 K,
T10 = 464 K, T20 = 546 K, and T75 = 753 K. Earth’s surface temperature of 287 K is
somewhat colder than that for one full layer (n = 1) at 303 K. The number of layers
for the Earth’s atmosphere (beyond the special zeroth layer) is obtained by solving
P1: GFZ
for n in Eq. 8.35, giving

n = (Ts /T0 )4 − 1 = (287 K/255 K)4 − 1 = 0.6. (8.36)
It is not surprising that Earth’s atmosphere contains only 60% of an IR layer since
O2 and N2 hardly absorb IR, leaving the task of IR absorption to trace amounts
of CO2 and H2 O. Venus, on the other hand, has a large temperature difference
between the upper atmosphere at T0 = 229 K and the surface at Ts = 750 K. These
temperatures give 100 IR layers for CO2 rich Venus!
8.4 Temperature Refinements

Better temperature estimates result when we consider additional energy paths
in the atmosphere in our calculation. Our next model uses three locations, two
atmospheric layers and Earth’s surface, along with these six additional energy
paths [K. Trenbeth, in IPCC’s Climate Change 1995]:
r Solar flux absorbed by the atmosphere, satmos = 67 W/m2
(70% in upper atmosphere, 30% in lower atmosphere).
r IR flux emitted from the surface directly to space, f IR = 40 W/m2 .
r Latent heat (evaporation) flux from the surface, f evap = 78 W/m2
(50% absorbed in upper and lower layers).
r Convective flux carried by air from the surface, f conv = 24 W/m2
(100% asorbed in lower atmosphere).
r Latent heat flux from evaporation of water is determined from global evaporation
of 5.2 × 1014 m3 /year (5.2 × 1017 kg/m3 -year):
Q̇ = ṁevap L evap = (5.2 × 1017 kg/m3 year)(1 year/3.2 × 107 s)(2.3 MJ/kg)
= 3.7 × 1016 W. (8.37)
The average latent heat flux is
f evap = (3.7 × 1016 W)/(4π )(6400 km)2 = 80 W/m2 . (8.38)
The heat balances for the upper surface of the atmosphere, the upper atmospheric
zone, and the lower atmospheric zone are given below:
(1 − a )so /4 = σ T 1 4 + f IR [top of atmosphere] (8.39)
2σ T 1 = σ T 2 + 0.7satmos + 0.5 f evap
4 4
[upper atmosphere] (8.40)
2σ T 2 4
= σ T 1 + σ T s − lIR + f conv + 0.5 f evap + 0.3satmos
4 4
[lower atmosphere]
(8.41)
These equations smooth the temperature difference from the surface to the upper
atmosphere, giving
r upper atmosphere T1 = 243 K
r lower atmosphere T2 = 272 K
r surface Ts = 287 K, which is the measured value.
P1: GFZ
8.4. Temperature Refinements 211
Figure 8.5. Multiple scattering of so-

lar flux between Earth’s surface and
clouds. (Adapted from J. Harte, Con-
sider a Spherical Cow, University Science
Books, 1988)
8.4.1 Lapse Rate

Atmospheric temperature drops as elevation increases with a lapse rate of
=
−6.5 K/km. The above result gave a lapse rate between the surface and lower
atmosphere if we assume the first layer is 3-km thick, giving
= (Ts − T1 )/g1 =
(287 K − 272 K)/3km = 5 K/km. The drop in temperature is caused by warm air
rising and expanding at reduced pressures, thus cooling the air. The dry adiabatic
lapse rate is derived under the assumptions of zero humidity and zero heat transfer
to rising air.4
8.4.2 Cloud Dynamics

Albedo is a measure of reflectivity from clouds and planetary surface. Albedo also
takes into account multiple optical bounces between the surface and clouds. An ac-
curate description of energy flow through clouds includes reflection, transmission,
and absorption in each microvolume as a function of depth in the cloud, scattering
angle and old/new photon energies. Measurements in 1995 showed that clouds
absorb 40% more than GCM calculations had predicted. We simplify clouds by
ignoring absorption and assign them a transmission coefficient T and a reflection
coefficient R, which conserve energy as T + R = 1. See Fig. 8.5.
4
Combine Q = 0 = c v T + pV (first law) and c p = c v + R with pV + Vp = RT
(perfect gas law for one mole) to obtain 0 = c v T + RT − Vp = c p T − Vp, and
T/p = V/c p = RT/c p p. Convert c p from J/mole-K to Cp in J/kg-K by multiplying c p by
molecular weight M in grams and dividing by 103 for kg, or T/p = V/Cp = 103 RT/Cp Mp.
The dry, adiabatic lapse rate from the chain rule is
d = ∂ T/∂z = (∂ T/∂ p)(∂ p/∂z). Using
pressure decrease with elevation z, ∂ p = −ρg∂z, and mass density of 1 mole (ρ = one mole
mass/V = Mp/103 RT), gives ∂ p/∂z = ρg = −(Mg/103 RT) p. This integrates to p = po e−z/H
with an effective height H = (103 RT/Mg) = 7.2 km. Finally, the dry adiabatic lapse rate is

d = ∂ T/∂z = (∂ T/∂ p)(∂ p/∂z) = −(103 RT/Cp Mp)(Mgp/103 RT)

d = −g/Cp = (10 m/sec2 )/(−1004 J/kg-K) = −10 K/km.
The measured lapse rate, −6.5 K/km, is smaller because additional heat is removed from
water vapor to lower temperature.
P1: GFZ
The flux reflected to space from clouds is Rc so , where Rc is cloud reflectivity

and so is incident solar flux. The solar flux transmitted through the clouds reflects
from Earth’s surface with a reflection coefficient Rs and is transmitted back through
the cloud to space with a flux TRs Tso . This process continues as the clouds reflect
Earth’s first reflection back to itself, allowing for a second reflection, which partially
passes through the cloud to space with a flux TRs Rc Rs Tso . The total flux sent to
space is
f = Rc so + T Rs Ts o + T Rs Rc Rs Ts o + T Rs Rc Rs Rc Rs Ts o + · · · (8.42)
This is an infinite sum, which we write as
f = Rc so + T Rs [1 + (Rc Rs )1 + (Rc Rs )2 + (Rc Rs )3 + · · ·]Ts o , (8.43)
or
∞
f = Rc so + i=0
(Rc Rs )i T 2 Rs so . (8.44)
Using the identity,

∞ 1
xi = , (x < 1). (8.45)
i=0 1−x
the last equation becomes
f = Rc so + T 2 Rs so /(1 − Rc Rs ) (8.46)
Let us compare the albedo a = f /so for a cloud above rock and for a cloud
above snow and ice. For the case of rock under a cloud (T = 0.7, Rc = 0.3,
Rs = 0.1), we obtain a = 0.35, slightly larger than the cloud at a = 0.30.
For reflecting snow at Rs = 0.6 under a cloud, we obtain a doubled albedo
a = 0.67.
8.4.3 High Clouds Versus Low Clouds

Cloud calculations are improved when cloud elevation is included, making a dis-
tinction between high and low clouds. High clouds reflect solar energy, thereby
cooling Earth beneath them. But since high clouds exist at lower temperatures,
their IR emissions are reduced. This traps heat, which on balance warms Earth. The
lower, warmer clouds also reflect energy, but they emit more IR since they are at
higher temperatures, which in the final analysis cool Earth. Thus, a main factor in
the debate on cloud effects hinges on the height and temperature of clouds: high
clouds warm while low clouds cool. Outgoing infrared along the equator varies
widely because of clouds. Over Indonesia, which has persistent high, cool clouds,
the IR flux is less than 200 W/m2 as compared to nearly 300 W/m2 over most of
the equatorial region.
P1: GFZ
8.5. Link Between CO2 and Temperature 213
8.5 Link Between CO2 and Temperature

8.5.1 General Circulation Models
GCM calculations can predict weather a week in advance. GCMs also predict sea-
sonal climatic changes and long-term scenarios, such as those driven by increasing
CO2 in the atmosphere. GCMs are based on physics, but parameters are added to
account for processes not explicitly resolved by physics. These parameters are ad-
justed to mirror improved measurements or new analysis of past climate records.
To some extent GCMs are similar to model simulations of nuclear weapon explo-
sions, which have adjustable parameters to fit results from nuclear tests. GCMs use
the following basic physical principles to obtain the results of Fig. 8.6:
Figure 8.6. US temperature changes simulated by GCM models. US projections are for the
lower 48 states based on historical and projected changes in atmospheric concentrations of
greenhouse gases and sulfate aerosols. In the twentieth century the models simulated a US
temperature rise of 0.7–1.9◦ F (0.4–1.1◦ C) compared to the observed range of 0.5–1.4◦ F (0.3–
0.8◦ C). The models projected a global temperature rise of 0.9–1.4◦ F (0.5–0.8◦ C), compared to
observations of 0.7–1.4◦ F (0.4–0.8◦ C). These GCM results are in reasonable agreement with
the data. For the 21st century, the models project global warming of 3–6◦ F (1.7–3.3◦ C) and
3–9◦ F (1.7–5.0 ◦ C) for the United States. The lower bound estimates assume lower emissions
of greenhouse gases than assumed by other models (US Global Climate Change Research
Program, 2000).
P1: GFZ
r Newton’s equations of motion

r moving frames with centripetal and coriolis accelerations
r fluids (turbulence, viscosity, hydrodynamics)
r gravitational and pressure gradient forces
r mass conservation (continuity equation)
r energy conservation
r equations of state
r water evaporation and condensation
r ocean-atmosphere coupling and ocean-current modeling
r surface features, albedo, ice/snow, salinity
r solar variation in time and location
r heat capacity and conduction
r gas, aerosol, and cloud properties.
GCMs typically use more than 20 layers to describe the atmosphere and more
than 10 layers to describe the oceans. Water’s higher viscosity gives smaller eddies,
forcing the use of smaller unit cells for ocean modeling as compared to atmospheric
models, which typically use 3000 cells of 200 km × 300 km. Surfaces are parame-
terized for ice flows and geographic coverings. The issue is not whether GCMs are
accurate but whether they are accurate enough to predict long-term climate changes.
8.5.2 Radiative Forcing

Additional CO2 raises atmospheric IR emission and absorption, elevating down-
ward heat flow from the atmosphere to the surface. This is seen in Eq. 8.28, which
gives increased surface temperature Ts with increased atmospheric emissivity ε a . It
is imperative to determine the amount of radiative-forcing flux f forcing (W/m2 ) from
concentration levels of greenhouse gases and other effects. Radiative forcing is the
extra downward flux of energy from greenhouse gases and other effects. Theory
and experiment show a connection between CO2 levels and radiative forcing,
f forcing = 3.75 ln[c/c o ]/ ln 2, (8.47)
where c is the concentration of CO2 and c o is the preindustrial level of 280 ppm.
Radiative forcing is, fortunately, not linear with CO2 concentration, but rather it
is logarithmically dependent. Thus, doubling CO2 gives f forcing = 3.75 W/m2 , but
quadrupling CO2 only doubles forcing to 7.5 W/m2 . Industrialization raised CO2
from 280 ppm to 370 ppm (2000), giving f forcing = 1.51 W/m2 . All other anthropogenic
gases and aerosols contribute additional radiative forcing [CH4 (+0.47), new H2 O
(+0.5), N2 O (+0.14), CFCs (+0.15), plus SF6 , O3 ], giving a total radiative forcing in
2003 of about 2.5–3.0 W/m2 .
The 1960s temperature lull is attributed to aerosols—micron-sized combustion
particles (sulfates and organic carbon). Aerosols reflect sunlight away from Earth,
reducing radiative forcing and increasing diffuse radiation. The inverse calculations
(based on fitting GCM results) for aerosol forcing give about −1 W/m2 , while for-
ward calculations (based on concentrations) give about −1.5 W/m2 . At any rate,
P1: GFZ
8.5. Link Between CO2 and Temperature 215
aerosols reduce temperatures below those values predicted from greenhouse gases
alone. So while this phenomenon has been used to explain the 1960s tempera-
ture lull, other factors, such as volcanic eruptions and, perhaps, solar variations,
are needed to explain the details of 20th-century temperature patterns. Nonethe-
less, aerosols are transient, with lifetimes of years, while CO2 has a lifetime of a
century.
Recall the base case, upper atmosphere temperature, calculated in Eq. 8.17 with-
out greenhouse gases and feedbacks,
Ta = [(1 − a )so /4σ ]1/4 = 255 K. (8.48)
For small changes in forcing compared to absorbed solar flux of 239 W/m2 , we
expect a linear increase in temperature T occurring with changes in forcing
f forcing . This linear coupling is defined as λ = dfforcing /dT, where coupling con-
stant λ is determined by taking the temperature differential of the radiation flux;
hence
λ = d f forcing /dT = d(σ T 4 )/dT = 4σ T a 3 = 4(5.6 × 10−8 )(255 K)3 = 3.7 W/m2 K.
(8.49)
IPCC projected that CO2 would rise to between 500 and 950 ppm by the year
2100. Doubling CO2 to 560 ppm gives f = 3.75 W/m2 and a conventional wisdom
temperature rise without feedback of
T = f forcing /λ = (3.75 W/m2 )/(3.7 W/m2 K) = 1 K. (8.50)
8.5.3 Feedback and T

The climate system can amplify or diminish forcing from greenhouse gases. For ex-
ample, enhanced CO2 warms Earth, reducing glaciers and polar caps. This replaces
white reflective surfaces with rocky or wet absorption surfaces, lowering albedo
and increasing solar absorption, which further raises Earth’s temperature. Simi-
larly, enhanced warming evaporates more water, increasing infrared absorption,
further raising the temperature, which evaporates more water. As indicated above,
high clouds raise the surface temperature (positive feedback) and low clouds lower
the surface temperature (negative feedback). Ocean thermal inertia lengthens time
scales, while unstable ocean currents reduce time scales. Frozen tundra can re-
lease methane (positive feedback), but it also can remove CO2 with photosynthesis
(negative feedback).
A doubled CO2 that comes with CH4 increases (1%/year) and N2 O increases
(0.3%/year) leads to forcing from CO2 (3.8 W/m2 ), CH4 (1 W/m2 ), H2 O (2.7 W/m2 ),
clouds (0.5 +/− 1 W/m2 ), ice (0.5 W/m2 ), and aerosols (−0.5 W/m2 ) for a total
of about 8 W/m2 . This gives a median temperature rise with feedback of Ts =
(8 W/m2 )/(3.7 W/m2 K) = 2.2 K, which is close to the IPCC projected median value
of 2.5 K with a range of 1.5–4.5 K (and a possible range of 1.4–5.8 K).
P1: GFZ
8.6 Solar and Oceanic Variations

8.6.1 Solar Variations
The Sun was 30% less intense billions of years ago with a flux so = (0.7)(1367 W/m2 )
= 957 W/m2 . A weaker Sun makes a colder Earth,
Ta = [(1 − a )so /4σ ]1/4 = [(0.7)(957)/4σ ]1/4 = 233 K, (8.51)
which is 22 K lower than today’s 255 K. The fractional reduction in temperature is
much less than the 25% solar reduction. We might expect solar variations of 0.2%
are possible since they are twice the present 11-year solar variation.5 This variation
could give extra solar forcing of
ssolar = 0.2%(1 − a )so /4 = (0.002)(0.70)(1365 W/m2 )/4 = 0.5 W/m2 , (8.52)
which is 6% of the projected rise of f forcing = 8 W/m2 from doubled CO2 with feed-
back. Most scientists reject the 0.2% solar variation as a cause of climate change
because temperature changes from T 4 radiation give small changes in high atmo-
spheric temperature. Twice the solar variation of 0.1% gives a temperature varia-
tion, using the differential of Eq. 8.17 divided by Eq. 8.17,
Ta = Ta (s/so )/4 = (255 K)(2 × 10−3 )/4 = 0.1 K. (8.53)
On the other hand, new data have caused scientists to refrain from ruling out a
solar contribution to climate change. A correlation was discovered in 1999 between
number of sunspots and surface temperature of Earth between 1870 and 1985. As
the number of sunspots increased from 30 to 90, surface temperature rose 0.5 K. This
might seem counterintuitive since colder “black” sunspots radiate less, but faculae,
which are bright patches on an active sun, far outshine dimmed sunspots. Earth had
unusually cold weather (1–1.5 K lower in winter than now) during the Maunder
minimum (1640–1720) when the peak number of sunspots was 30, as compared
to 90 in warmer times. This implies that Sun’s intensity was reduced by 0.25%.
A better correlation was obtained between sunspot cycle length and temperature
anomaly. A shorter cycle-time implies a more active Sun. As the cycle shortened
from 11.7 to 10.3 years, Earth warmed by 0.5 K, 5 times our 0.1 K estimate.
If it is true that solar variations contribute to climate change, then what mech-
anism makes this happen beyond that described by our simple theory? GCM cal-
culations show extra heating in summer, warming the stratosphere, strengthen-
ing easterly winds and changing wind patterns. However, the GCM changes are
smaller than the observed changes. Other GCM calculations, which include interac-
tive stratospheric chemistry with ozone, had more success in predicting an 11-year
climate cycle. A theoretical link needs a more active sun to emit 10 times more ul-
traviolet. Extra UV would interact with ozone, raising stratosphere temperatures,
5
Main sequence stars, HD 76572 and HD 81809, similar to the sun in size and age, have flux
changes of 0.23% and 0.42%, respectively.
P1: GFZ
8.6. Solar and Oceanic Variations 217
but this would only raise the surface at high latitudes by only a few tenths of a
degree.
8.6.2 Rising Oceans

Oceans are rising 1.5–2.0 mm/year (15–20 cm/century). Thermal expansion of the
warming oceans contributes about 0.5 mm/year, with the rest of the rise due to
melting ice. This rate was exceeded in a given period 14,000 years ago, when the
oceans rose 4 m per century, rising 20 m in 500 years.6 GCM calculations show
that global changes of 0.3 K can cause larger regional changes of 1–2 K in the
Northern Hemisphere by changing the Arctic/North Atlantic ocean oscillations. A
global temperature rise of 2.5 K would cause a 5 K rise in the polar regions. As the
North Polar Cap melts, ocean levels do not rise since the cap is floating ice. Some
conjecture that the oceans could rise faster and higher should the West Antarctica
Ice Sheet (WAIS) dislodge catastrophically, creating a gigantic tsunamis. Others say
WAIS is stable since it was created 10 million years ago and has been shrinking for
some time. A launched WAIS would dramatically raise levels because of its being
grounded on the sea floor. The volume of WAIS is
VWAIS = (1.5 × 106 km2 area)(1.5 km thickness) = 2.2 × 106 km3 , (8.54)
which would be reduced as it melted by 10% to 2.0 × 106 km3 . Spreading this
volume over all oceans raises their height by
VWAIS /Aocean = (2.0 × 106 km3 )/(71%)(4π )(6400 km)2 = 5.5 m, (8.55)
enough to cover 30% of Florida and Louisiana. The complete melting of Antarctica
and Greenland ice is unlikely, but should it occur it would raise oceans by 100 m,
destroying most of Earth’s major cities. See Fig. 8.7 for estimates of level rise.
8.6.3 Unstable Ocean Currents and Rapid Climate Change

Deep ocean currents move heat from hotter low latitudes to colder high latitudes
(Fig. 8.8). The winds, such as the El Nino southern oscillation, interact with ocean
currents to periodically change climate patterns. There is evidence that climate
change can destabilize deep ocean currents. An examination of data on salinity,
phosphate, silica, temperature, and isotope ratios reveals a pattern of past insta-
bilities in ocean currents. At one point the temperature rose 10◦ C in less than a
decade when the currents were disrupted. Wallace Broecker of Columbia Univer-
sity postulated that 8200 years ago during that period of warming, fresh water
from melting 3-km thick glaciers and excessive rain flowed from the large Lake
Agasswiz in Canada down the St. Lawrence Seaway in to the North Atlantic. The
lighter, fresh water floated on top of the north-flowing, warm, high-saline water.
6
The Dutch are concerned about rising oceans, especially after the 1953 flood which killed
2000 and left 100,000 homeless over 15% of the Netherlands.
P1: GFZ
Figure 8.7. Sea level rise projections. All three projections consider thermal expansion of
the oceans, while the middle curve also considers glacial melting. These estimates do con-
sider polar melting or snow accumulation on Greenland and Antarctica (US Global Climate
Change Program, 2000).
Figure 8.8. Ocean circulation conveyor belt. This simplified drawing shows the ocean con-
veyor belt, carrying heat from warm regions to cold regions (US Global Climate Change
Research Program, 2000).
P1: GFZ
8.6. Solar and Oceanic Variations 219
Figure 8.9. Abrupt climate change. Evidence of temperature changes of 5–10 K that occurred
over a few decades, inferred from measurements of 18 O/16 O isotope ratios. [ J. Severinghaus,
USGCRP, 1998]
Normally the higher density of the colder, saline polar water is sufficient to drive
the conveyor of current, but fresh water displaced its position on the surface and
stopped the normal currents. The current flow is large, equal to that of 100 Amazon
Rivers, which is 20 Sverdrups, or 20 × 106 m3 /s. If heat is denied to the upper lat-
itudes, they become colder and lower latitudes become warmer. Thus, a possible
unpleasant byproduct of climate change could be a frozen northern Europe and ex-
cessively hot tropics. These effects would be accompanied with a positive albedo
feedback, as more northern ice reflects more sunlight away from Earth, causing
further cooling. These views are not well understood by the wider community of
natural scientists, but they have been strongly supported by a panel of the National
Academy of Sciences (2002). See Fig. 8.9.
About 25% of transported heat moves to high latitudes by ocean currents, while
the remaining 75% is moved by atmospheric currents. Heat power contained in a
4-K warmed North Atlantic Deep Water (NADW) is
dQ/dt = (dM/dt) cT = (2 × 107 m3/s)(4200 J/kg K)(103 kg/m3 )(4 K) = 4 × 1014 W.
(8.56)
This is 20% of the average solar power absorbed by the north Atlantic,
d Q/dt = savg A = (200 W/m2 )(1013 m2 ) = 2 × 1015 W. (8.57)
As shown earlier, global latent evaporation power is 80 W/m2 , which evaporates

1 m/year. This is 40% of the absorbed 200 W/m2 , with the remaining 60% split
between NADW and atmospheric currents.
P1: GFZ
8.6.4 Ocean Acoustical Tomography

The ocean can act as a thermometer, measuring its own warming. Here is how: Light
is trapped inside fiber optic cables, since the speed of light in the cable is smaller
than the speed of light outside the cable. For the same reason, ocean acoustic waves
are constrained to a channel between the ocean surface and a 1-km depth. Wave
velocity is a minimum at a 1-km depth. The minimum results from a competition
between increased pressure with depth (which increases speed by 1.5 m/s per 100 m
of depth) and lower temperatures with depth (which reduces speed by 4.6 m/s ◦ C).
Pressure eventually wins the race since it continues to increase with depth, while
temperature is constrained to a drop of 20◦ C. This competition gives larger speeds
at depths below 1 km.
Time delays are measured with vertical arrays that can be inverted to obtain
three-dimensional temperature maps of the ocean. The speed of sound in saltwater
varies from 1440 m/s near Antarctica to 1520 m/s in the Mediterranean Sea. Using
an average speed of 1500 m/s, the time for a 10,000-km journey is
t = distance/velocity = (107 m)/(1500 m/s) = 6700 s = 2 h. (8.58)
If the oceans continue to warm at the present rate of 0.1–0.2◦ C in three

decades (0.004◦ C/year), the annual increase in sound wave velocity will be
(0.004◦ C)(5 m/s ◦ C) = 0.02 m/s. This shortens transit time by a measurable

v
t = t = (6700 s)(0.02 m/s)/(1500 m/s) = 0.1 s. (8.59)
v
However, time delay is an integral effect, manifesting many velocities along its
path. By combining satellite altimeter data for thermal expansion and acoustical
tomography data from time delays, the Acoustic Ocean Thermometry Consor-
tium (ATOC) measured an annual (summer-to-winter) flux oscillation of 150 +/−
25 W/m2 . This result is expected from solar variation over the seasons, as the fol-
lowing calculations demonstrate. About 70% of the total transmitted solar flux is
absorbed by Earth’s surface, or 0.7(0.7 × 1367 W/m2 /4) = 170 W/m2 . Peak hori-
zontal solar flux in W/m2 at noon (snoon ) at 30◦ N for the three seasons gives the
ratio (Section 12.1)
snoon (sum/equi/win) = 1.1/0.88/0.52. (8.60)
The length of a solar day at 30◦ latitude is 14 h at summer solstice, 12 h at the two
equinoxes, and 10 h at winter solstice. The daily integrated solar energy per day
is proportional to the length of the solar day times the peak flux at solar noon, or
QαTday snoon , giving a ratio of daily integrated solar energies Q for the three seasons,
Qsum /Qequi /Qwin = (14 × 1.1)/(12 × 0.88)/(10 × 0.52) = 15.4/10.6/5.2. (8.61)
We normalize this ratio to 1.0 for an equinox day,
Qsum /Qequi /Qwin = 1.46/1.0/0.5. (8.62)

P1: GFZ
8.7. Heat Islands 221
Hence, Earth at 30◦ N latitude receives three times more energy per unit area in
summer than in winter. Using this ratio, the average solar flux absorbed in summer
is ssum = 1.46 × 170 = 250 W/m2 and in winter it is swin = 0.52 × 171 = 90 W/m2 .
The difference between summer and winter is
s = ssum − swin = 250 W/m2 − 90 W/m2 = 160 W/m2 , (8.63)
in agreement with the measured summer–winter difference of 150 +/− 25 W/m2 .
Here is how we use the ocean as a thermometer: It can be concluded that this
difference flux causes a seasonal temperature change of 2◦ C over the top 250 m of
the ocean.7 The extra energy stored under 1-m2 area in the summer season is
Qsum-season = VρcT = (250 m3 )(103 kg/m3 )(4200 J/kg ◦ C)(2◦ C) = 2.1 × 109 J/m2 .
(8.64)
This is similar to the integrated extra seasonal heat over a 6-month period,
T/4
Qsum-season = (s/2)[1 + sin(2π t/T)]dt = (s/2)(T/2). (8.65)
−T/4
Integration time is 6 months with T = 1 year, giving

Qsum-season = (s/2)(T/2) = (75 W/m2 )(1.6 × 107 s) = 1.2 × 109 J/m2 . (8.66)
The oscillation amplitude s/2 = 75 W/m is one-half the summer–winter differ-
2
ence of s = 150 W/m2 .
8.7 Heat Islands

Dark asphalt roofs with 4% reflectivity can be 50◦ C (90◦ F) hotter than ambient
temperature during summer, while light-colored roofs with 80% reflectivity might
rise only 10◦ C (18◦ F). This effect makes urban areas 4◦ C (8◦ F) warmer than neigh-
boring countryside, increasing the energy load for air conditioning, which in turn
increases pollution. In the 1930s, the Los Angeles basin had considerable irrigated
orchards. In 1934, the high LA temperature was 97◦ F, but in the 1990s it rose to over
105◦ F. Higher temperatures require more air conditioning, as electrical power rises
nearly 2% for each 1◦ F rise. Higher temperatures also create more smog, as smog
incidences increase by 3% for every 1◦ F rise. See Fig. 8.10.
Planting more trees helps. A mature, watered 30-foot tree evapotranspires 40
gal of water a day. This cooling is equivalent to absorbing the heat from a small
space heater run for 4 h. To counter the effect of heat islands, Arthur Rosenfeld,
California Energy Commissioner, suggested plans to use reflective roadways and
roofing and to use more trees. Such a plan could save Los Angeles $1 billion/year
in energy costs. One way to make this happen is to use Clean Air Act provisions
7
A small portion of this power, 0.44 W/m2 , is retained beyond the seasonal oscillation to
slowly warm the oceans. (Levitus, et al., 2001).
P1: GFZ
Figure 8.10. Los Angeles heat island. LA high annual temperatures after 1880 were partially
driven by aerosol coming from the 1883 Krakatau eruption, while temperature rise after 1935
was caused by urbanization. [A. Rosenfeld, California Energy Commission]
for commercial trading of pollution credits between industries in LA (Sectoin 6.2).

If industry A needs to buy an expensive mitigation measure to reduce pollution, it
could instead pay industry B to mitigate its plant’s pollution more effectively. Such
an approach would include reflective surfaces as a salable pollution credit.
Heat island effects are exaggerated if internal heat is directly added to solar flux.
For example, New York City’s area A = 830 km2 has a population of 8 million,
but commuters and tourists easily raise the effective population to over 10 million.
If each person’s thermal power consumption were the national average of 12 kW,
New York would have a thermal power of Pt = 1.2 × 1011 W and an internal heat
flux of
f int = Pt /A = (1.2 × 1011 W)/(8.3 × 108 m2 ) = 145 W/m2 . (8.67)
New York’s human energy flux is similar to Earth’s average absorbed solar flux
of 171 W/m2 (Section 8.3). The heat balance at the top of the atmosphere for a world
with New York’s power density is
(1 − a )so /4 + f int = 239 W/m2 + 145 W/m2 = σ T a 4 . (8.68)
This gives an upper atmospheric Ta = 287 K, 32 K higher than the present value of
255 K. For an atmosphere of n = 0.6 layer (Eq. 8.36), the surface temperature is
Ts = (0.6 + 1)1/4 Ta = (1.12)(287 K) = 321 K, (8.69)
34 K higher than the average surface temperature of 287 K. Thus, adding internal
heat to solar flux exacerbates heat islands.
P1: GFZ
8.7. Heat Islands 223
A better approach for heat island heat transfer is to use a box model, similar to
that used for air pollution (Section 6.3). Warm air, contained in a box of inversion
height H and area A, is blown out of the box with wind velocity u through a side
of length L wind . New York’s heat output is the product of its internal flux and its
surface area. In the steady state, the heat output of New York is balanced by wind
carrying warm air through the area of one side of the box,
Pt = f int A = 1.2 × 1011 W = (uρc p T)(H L wind ). (8.70)
This gives a temperature rise of T = 2◦ C, using L wind = 30 km (in a direction

perpendicular to the wind), H = 300 m, u = 5 m/s (11 mph), ρ = 1.29 kg/m3 , and
c p = 1000 J/kg ◦ C. If wind velocity is cut in half, T is doubled to 4◦ C. Tempera-
ture is further increased if the wind is parallel to the length of Manhattan Island,
which means more time is needed to remove heat. Smog further exacerbates the
situation, since dirty air strongly traps infrared and it absorbs more solar energy.
In the box model, a 20% reduction in absorbed heat reduces temperature eleva-
tion by 20%. This could be accomplished by reducing energy consumption or by
reducing absorbed solar energy with reflecting streets/roofs and with more trees.
California adopted standards and funding to take advantage of white roofs, an
approach which DOE estimated could save the entire US 7 GWe of summertime
peak power. See Fig. 8.11 for reflectivity of roofing materials and Fig. 8.12 for Los
Angeles temperature data.
Figure 8.11. Solar reflectivity versus roof temperature. By increasing roof reflectivity from
4% to 80%, roof temperature drops as much as 40◦ C. [A. Rosenfeld, Lawrence Berkeley
Laboratory]
P1: GFZ
Figure 8.12. Ozone, peak power, and temperature in central LA. A linear correlation is
observed between 4 PM downtown temperature and the peak power use for air conditioning.
The additional power also raises smog levels. [A. Rosenfeld LBL]
8.8 Policy Options

This section describes options for reducing carbon with technologies that en-
hance end-use efficiency and produce electricity more efficiently. In addition,
policymakers can consider sequestering CO2 , establishing a carbon tax, trading
carbon allotments, and setting national carbon limits.
8.8.1 CO2 Reductions

Carbon emissions from coal plants (0.24 kg C/kWh) could be reduced by 60% by
substituting combined-cycle gas turbines (CCGT, 0.1 kg C/kWh) fueled with natural
gas. The reduction is the result of two factors: CCGTs high efficiency (55–60%) in
converting energy to electricity as compared to coal (35–40%) and natural gas’s
energy-laden hydrogen, which does not make CO2 .
P1: GFZ
8.8. Policy Options 225
Enhanced end-use efficiency greatly reduces carbon consumption. This approach

was examined in the 1997 DOE Five-Lab study (Chapter 14), which concluded
that 0.4 Gt/year of carbon could be saved on a cost-effective basis by improving
buildings, transportation, industry, and utilities. The Five-Lab Group projected a
rise from today’s 1.550 Gt/year of carbon to 1.922 Gt/yr in 2020 under a business-
as-usual consumption scenario. They projected a 9% reduction to 1.743 Gt/year
with moderate policy actions, a figure that could be further reduced (23%) to 1.478
Gt/year with a $25/tonne carbon tax or lowered even more (29%) to 1.357 Gt/year
with a $50/tonne carbon tax.
8.8.2 Sequestered CO2

To sequester CO2 is to store it outside the atmosphere. For carbon storage to be ef-
fective, it must not have a large leakage rate. At today’s carbon consumption rate,
650 Gt of carbon will be produced in the 21st century. If storage had 1%/year leak-
age, the leakage rate would equal today’s carbon use rate. Underground storage of
CO2 is currently being used to force additional petroleum to the surface in Texas and
Colorado at a cost of $10–15/ton of CO2 (0.5–1 c/kWh). Carbon dioxide is favored
because it is miscible at 1000 psi and 31◦ C as compared to nitrogen at 5000 psi and
240◦ C. In the United States, sequestered CO2 is obtained as a by-product of natural
gas production, and not directly from combustion. This sequestering is not per-
manent because the CO2 is used in the oil industry. Similarly, Norway sequesters
1 Mton CO2 /year in the North Sea, responding to a government incentive to fill
depleted natural gas wells. However, depleted oil and gas wells have a limited
volume and would be filled at some point. The ocean is probably the most likely,
long-term storage location for CO2 .
Experiments in Monterey Bay, CA, sequestered liquid CO2 at a 3.6-km ocean
depth, where liquid CO2 is denser than water. Under these conditions CO2 natu-
rally combines with water to form the hydrate CO.2 6H2 O, which dissolves into the
ocean. Residency time for sequestered CO2 should be long since the oceans are un-
saturated, but costs and impacts are uncertain. Another approach is the Subarctic
Pacific Iron Experiment for Ecosystem Dynamics Study in which iron in FeSO4 is
dissolved in the ocean to sequester CO2 . In the experiment, 350 kg of iron increased
photoplankton and decreased dissolved CO2 over a 100-km2 area. However, con-
tinual seeding of all the oceans would be difficult. Perhaps the best approach is
similar to natural weathering, which combines CO2 with CaCO3 or CaSiO3 to form
HCO− 3 . The world’s rivers now transport 1 Gt/year carbon to the oceans, some of
which sinks to the bottom after biological extraction.
Any plan to sequestering a large fraction of the global rate of 6.5 Gt C/year would
be daunting. To put things in perspective, let us discuss this in terms of freight
trains. A 1 GWe power plant consumes a coal train loaded with 104 tons/day,
which is transported to the plant by 100 freight cars carrying 100 tons/car. A global
consumption of 6.5 Gt/year of carbon corresponds to a coal train convoy of
(6.5 × 109 ton/year)(1 year/365 days)/(104 ton/train) = 2000 trains/day. (8.71)

P1: GFZ
This volume is increased by a factor of 1000 as it is converted from solid hydrocarbon

to gaseous CO2 .
8.8.3 Fractional Carbon Tax

If carbon taxes are small, it is possible to make estimates of carbon consumption
using the elasticity of demand,
d/d
ed = , (8.72)
p/ p
The fractional reduction in demand, d/d, is proportional to the fractional in-
crease in price, p/ p, with the demand elasticity acting as the proportionality
constant. As a beginning, we assume all carbon prices are taxed at a fixed rate of
20%, or p/ p = +0.2. Choosing a value for global carbon demand elasticity is dif-
ficult, since there is a dearth of specific global data. Nonetheless, from experience
with the US gasoline market we choose e d = −0.2. (In fact, e d is not a constant, but it
varies with price, with the amount of price rise, with the type of fuel, with the local
economic situation, and with other factors.) Thus, demand elasticity is a function
and not a single parameter as it varies from fuel to fuel and from low prices to
high prices. A 20% global tax would be about $0.50/gal in the United States, but in
Europe and Japan it would be $0.80/gal. Nonetheless, the global reduction in carbon
from a 20% carbon tax would be about
d = −e d (p/ p)d = (−0.2)(0.2)(6.5 × 109 ton/year) = −0.25 Gton/year. (8.73)
This is a 4% reduction of global consumption at e D = −0.2 and an 8% reduction
at e D = −0.4. Such a tax could be considered politically Draconian since it varies
between countries, giving a lower tax to some of the nations that consume the most.
8.8.4 Carbon Mass Tax

An alternative approach is a mass-based carbon tax. We begin with the upper range
under discussion of $50/ton. Since petroleum is 88% carbon by weight, the tax
would be ($50/ton C)(0.88) = $40/ton on petroleum. A gallon of petroleum has a
3.5-kg mass, which would be taxed at
($40/ton)(1 ton/1000 kg)(3.5 kg/gal) = $0.14/gal. (8.74)
Such a tax could raise a great deal of revenue for the United States:
(20 Mbbl/day)(42 gal/bbl)($0.14/gal)(365 day/year) = $43 billion/year. (8.75)
The DOE Five-Lab study estimated that a $50/ton tax, plus aggressive poli-
cies, could cut 2020 emissions by 0.4 Gt/year, a 20% reduction of a US-projected
2 Gt/year (Fig. 8.13). Does this reduction come from market forces alone? If market
forces are strong, we say the market is elastic with e d close to −1. If society does not care
about price, then we say markets are inelastic with e d closer to zero. The Five-Lab re-
sult implies an average elasticity e d = (d/d)/(p/ p) = −0.2/($0.14/$2.50) = −3.
P1: GFZ
8.8. Policy Options 227
Emissions
(MiC)
2100
Business-As-Usual
Scenario
1900
ential
Resid
ercial
Comm
Industrial
1700
Transportation
1500 Electric Utilities
Advanced Scenario
1300
1990 1995 2000 2005 2010 2015 2020
Figure 8.13. US Carbon emission reductions by sector with a carbon tax of $50/tonne (Brown
et al., 1998).
Since energy markets are not superelastic, we see the projection depends more on
the adoption of energy-efficient technologies than a hoped-for reduced consump-
tion from direct pricing. This is what the DOE study also concluded.
8.8.5 Btu/Joule Tax

This approach is similar to the mass tax, but it is a tax on energy content. A Btu tax
of 10% on US gasoline amounts to
(0.1)($2.50/gal)/(130 MJ/gal) = $2/GJ = $2/MBtu. (8.76)
8.8.6 The Kyoto Protocol

The 1997 Kyoto Protocol to the UN Framework Convention on Climate entered
into force in 2005. It limits six gases (CO2 , CH4 , N2 O, and three CFCs) to a level
that is effectively 5.2% below the 1990 level. The first compliance period is from
2008 to 2012, with developed nations taking larger cuts (European Union, 8%; the
United States, 7%; and Japan, 6%). During the 1990–95 period, the EU dropped by
1% below the limit, while the United States and Japan both grew by 7%. In 2001,
President George W. Bush withdrew the United States from the protocol, stating
that developing nations, such as India and China, were not covered, in spite of
the fact that they were expected to consume much more carbon in the future. On
P1: GFZ
Figure 8.14. Worldwide emissions of carbon by region. [US Carbon Dioxide Information
Analysis Center, Oak Ridge National Laboratory]
the other hand, to comply with the Kyoto Protocol the United States would have
to cut 7% of its 2001 total of 1.56 Gt/year, in addition to cutting the effects from
population growth. If US energy growth followed its 0.9%/year population growth,
the required reduction over 15 years to the 1990 level would be
(1.00915 − 1)(100%) + 7% = 14% + 7% = 21%, (8.77)
and 27% over 20 years. In July 2001, 178 of the 179 nations (the US absent) agreed
to the Koyto–Bonn Accord, which moves the Kyoto Protocol closer to implemen-
tation through the adoption of carbon-allowance trading. The compromise gave
the United States 55 tons/year exemption since US forests sequester CO2 . Without
the United States, the new accord would only reduce the 1990 base-year carbon
emissions by 2% by 2050, whereas the original Kyoto Protocol (with the United
States) would have cut emissions by 15% by 2050. The European Union ratified the
Kyoto–Bonn Protocol in February 2002, making EU liable in court if its countries
fail to make their carbon limits. See Fig. 8.14.
The Kyoto Conference chose national carbon limits and rejected mandatory taxes.
The Protocol creates mechanisms to allow nations to trade greenhouse gas al-
lowances, similar to the US practice of trading sulfate allotments under the Clean
Air Act (Section 6.2). This approach allows countries to achieve reductions at lower
costs by purchasing less expensive allotments from countries that have achieved
their targets or that can achieve them at a cheaper cost. With the United States
not buying carbon allotments, allotment prices in 2010 are estimated to drop from
$50/ton to $15/ton. (In 2005 the allotments were $10/ton.) Cost estimates for US
implementation of the Protocol varied, from 0.1% of GNP (high adoption of energy
efficient technologies) to 2% of GNP (if no savings from international emissions
trading). These percentages translate to annual costs between $100 to $1000 per
household.
P1: GFZ
Problems 229
Figure 8.15. Carbon dioxide emissions per unit of gross domestic product. Data from 1990–
2002 and projections from 2002–2025 in tons of carbon equivalent per million 1966 dollars
of GDP. [Annual Energy Outlook 2003, EIA]
In February 2002, President Bush proposed a voluntary 18% carbon reduction by

2012 in terms of greenhouse gas intensity, which is the ratio of national total energy use
of carbon fuels to GDP dollar. This reduction is a voluntary limit and not a legal cap;
it is stimulated by $4.6 billion in tax credits for renewable energy. Greenhouse gas
intensity has been naturally falling over the years, from about 19,000 Btu/$ in 1970
to 10,600 Btu/$ in 2000, at a rate of 2%/year. The Bush proposal merely maintains
this trend with an 18% reduction in 10 years from 183 ton/$1M to 151 ton/$1M.
This 2%/year reduction in greenhouse gas intensity will not reduce US carbon
emissions, since the GDP continues to grow and there are no legal limits. This is
readily seen with the EIA data and projections shown in Fig. 8.15.
Problems
8.1 Global carbon. 1999 global coal consumption was 4.7 billion short tons of coal,
of petroleum was 74.9 Mbbl/day, and of natural gas was 84 TCF. (a) How many
tons of carbon were consumed from coal (80% C), petroleum (average CH1.5
at 140 kg/barrel), and natural gas (CH4 )? What fraction of CO2 remained in
the atmosphere if it rose 1.5 ppm in 1999?
8.2 Your carbon. (a) How much carbon do you emit driving 10,000 miles a year
at 28 miles/gal? (b) How much carbon do you alone emit if the US popu-
lation of 280 million consumes, for simplicity, 85 quads/year of iso-octane,
C8 H18 ?
8.3 Carbon sinks. What is the ratio of carbon in the atmosphere (1999, 370 ppm)
to that in global forests (15 kg C/m2 on 50 × 1012 m2 )?
8.4 Future CO2 scenarios. Modify Eq. 8.13 for integrating atmospheric carbon by
using the CO2 100-year atmospheric mean lifetime?
P1: GFZ
8.5 Human heat. (a) How does human usage of 382 quads/year (1999) affect the
255 K upper atmosphere temperature?
8.6 Summer/winter. Earth is 3% closer to the Sun in winter as compared to sum-
mer. (a) How much does this shift the Ta of 255 K between seasons? (b)
Why does the northern hemisphere have more severe winters than the south-
ern hemisphere, in spite of the fact that the Sun is closest in the northern
winter?
8.7 Differential changes. Obtain the differential form of Earth’s radiation balance
(Eq. 8.17) using fractional changes, such as a /a for the fractional change in
albedo. What is the change in Ta if the solar constant increases by 10% and the
albedo rises from 0.30 to 0.33.
8.8 Saturn’s temperature. What is Ta for Saturn at 10 astronomical units with a
cloudy atmosphere of 0.7 albedo?
8.9 Energy Branching Ratios. How do the assumptions leading up to Eqs. 8.39–
8.41 compare to the actual energy branching ratios in Fig. 8.4?
8.10 Evaporation power. (a) What thermal power is needed to evaporate rainfall
of 1 m/year? (b) How does this compare to the power needed to raise this
water vapor 2 km?
8.11 Lapse rate. What water content explains the reduction of the dry lapse rate
from −10 K/km to the measured rate −6.5 K/km?
8.12 Albedo. (a) What is Earth’s net albedo when it is 50% cloudy (40% transmis-
sion, 60% reflection) with 10% surface reflection. (b) Modify the theory to take
into account cloud absorption.
8.13 Burn all coal. (a) How many parts per million of CO2 result from 20–100%
burning of global coal reserves of 1016 kg (80% C). (b) What is the radia-
tive forcing (W/m2 ) from this, and what is the temperature rise without
feedbacks?
8.14 Old records. Analysis of the Vostok ice shows that methane and CO2 concen-
trations in historic air samples correlate with local temperatures over the past
160,000 years. At one point, CO2 concentration dropped from 280 to 190 ppm,
while the temperature dropped 10◦ C. How does this compare to the text equa-
tions?
8.15 Solar coupling. Some speculate that solar faculae (bright spots) give ex-
tra ultraviolet that is absorbed by ozone in the stratosphere, lowering
albedo, giving more energy to Earth. How much albedo reduction raises Ta
by 2 K?
8.16 Rising oceans. (a) How much would oceans rise from melting all Greenland
(14 million km3 ) and Antarctica (30 million km3 )? (b) How much heat would
be needed to do this? How long would it take to melt the ice with 1 W/m2 ? (c)
Library question: What change in Earnest Shakelton’s ship Endurance would
have saved it in the Weddell Sea, 1915?
8.17 Warmed oceans. How much would the ocean rise if the temperature of the top
1 km rose 0.1◦ C (the present situation) or 2.5–5◦ C in the future? Water density
is 999.13 kg/m3 at 15◦ C and 997.07 kg/m3 at 25◦ C.
P1: GFZ
Bibliography 231
8.18 Heat islands. Black asphalt and rooftops absorb on a total basis an extra 10% of
incident solar energy. What is the temperature rise in a city (40 km × 40 km)
with a 500-m inversion height in a 5 m/s wind?
8.19 Carbon sequestering. What is the STP volume/year of CO2 from US con-
sumption of 1.5 Gt/year carbon?
Bibliography
Bender, M., M. Battle and R. Keeling (1998). The oxygen balance of the atmosphere, Ann.
Rev. Energy Environ. 23, 207–223.
Brown, M., et al. (1998). Engineering-economic studies of energy technologies to reduce
greenhouse gas emissions, Five-Lab DOE Study, Ann. Rev. Energy Environ. 23, 287–386.
Ehrlich, P. (1996). Disease and global circulation models, Ann. Rev. Energy Environ. 21, 125–
144.
Firor, J. (Ed.) (1995). Global Warming: Selected Reprints, American Association of Physics
Teachers, College Park, MD.
Hansen, J., et al. (2000). Global warming in the twenty-first century: An alternate scenario,
Proc. Natl. Acad. Sci. 97, 9875–9880.
Harte, J. (1988). Consider a Spherical Cow, University Science Books, Sausalito, CA.
———(1997). Carbon dioxide feedbacks, Ann. Rev. Energy Environ. 22, 75–118.
———(2000). Consider a Cylindrical Cow, University Science Books, Sausalito, CA.
Harvey, L. (2000). Global Warming: The Hard Science, Prentice Hall, Englewood Cliffs, NJ.
Herzog, H. and E. Drake (1996). Carbon dioxide recovery and disposal from large energy
systems, Ann. Rev. Energy Environ. 21, 145–166.
Holloway, S. (2001). Storage of fossil fuel-derived carbon dioxide beneath the surface of the
Earth, Ann. Rev. Energy Environ. 26, 145–166.
Intergovernmental Panel Climate Change (1994). Radiative Forcing of Climate Change, Cam-
bridge Univ. Press, Cambridge, UK.
———(1997). An Introduction to Simple Climate Models Used in the IPCC Second Assessment,
Report, Cambridge Univ. Press, Cambridge, UK.
———(2000). Climate Change 2001: The Scientific Basis, Cambridge Univ. Press, Cambridge,
UK.
———(2000). Emission Scenarios, Cambridge Univ. Press, Cambridge, UK.
Levi, B., D. Hafemeister and R. Scribner (Eds.) (1991). Global Warming: Physics and Facts,
American Institute of Physics Press, New York.
Levitus, S., J. I. Antonov, J. Wang, et al. (2001). Anthropogenic warming of the Earth’s climate
system, Science, 292, 267–270.
Melillo, J., et al. (1996). Tropical deforestation and the global carbon dioxide budge, Ann.
National Research Council (1991). Policy Implications of Greenhouse Warming, National
Academy. Press, Washington, DC.
———(1994). Solar Influence on Global Change, National Academy Press, Washington, DC.
———(2001). Climate Change Science: An Analysis of Some Key Questions, National Academy
-Office Technology Assessment (1991). Changing by Degrees: Steps to Reduce Greenhouse Gases,
OTA, Washington, DC.
P1: GFZ
Parson, E. and K. Fisher-Vanden (1997). Integrated assessment of global climate change,

Ann. Rev. Energy Environ. 22, 589–628.
Peixoto, J. and A. Oort (1992). Physics of Climate, American Institute of Physics Press, New
York.
Rosenfeld, A., et al. (2000). Technologies to reduce carbon dioxide emissions in the next
decade, Phys. Today 53(11), 29–34.
Schipper, L., et al. (2001). Indicators of energy use and carbon emissions, Ann. Rev. Energy
Environ. 26, 49–81.
Trenbeth, K. (1996). IPCC’s Climate Change 1995, Cambridge Univ. Press, Cambridge, MA.
US Energy Information Administration (1998). Impacts of the Kyoto Protocol on US Energy
Markets and Economic Activity, EIA, Washington, DC.
———(1999). Emissions of Greenhouse Gases in the United States, EIA, Washington, DC.
US Global Climate Change Research Program (2000). Our Changing Planet, USGCRP, Wash-
ington, DC.
———(2000). Climate Change Impacts on the United States, Cambridge Univ. Press, Cambridge,
UK.
Weart, S. (2003). The discovery of rapid climate change, Phys. Today 55(8), 30–36.
Wigley, T., and D. Schimel (Eds.). The Carbon Cycle, Cambridge University Press, Cambridge,
UK.
Zebrowski, E. (1997). Perils of a Restless Planet, Cambridge Univ. Press, Cambridge, UK.
P1: GFZ
9
Electromagnetic Fields and
Epidemiology
9.1 Power Line Health Effects?

This chapter examines possible health threats from power line electromagnetic
fields (EMFs) by examining physical mechanisms that might cause cancer. This is fol-
lowed with epidemiology approaches to show how statistical connections between
health effects and pollution exposure might be established.
A 1979 report by Nancy Wertheimer and Ed Leeper stated that enhanced levels of
childhood leukemia (a rare disease) were found in populations living near electrical
power lines.1 In spite of the fact that this study was widely criticized, it launched
lawsuits and new standards that cost billions of dollars. In response to the litigation,
many disciplines carried out wide-ranging research to examine the likelihood of
a link between cancer and EMFs at extremely low frequencies (ELF). In 1996, a
National Academy of Sciences panel concluded that essentially none of the 60-Hz
ELF-EMF data showed a viable link to cancer, except a statistically weak association
(or correlation) between incidence of childhood leukemia and populations living
near power lines, which was quantified with wire-code ratings. The NAS panel
further concluded that an apparently weak association with power line fields (wire-
code ratings) was not meaningful since the measured magnetic field exposures in
these homes did not correlate with childhood leukemia. In 1997, a National Cancer
Institute epidemiology study found “no association between the risk of [childhood
leukemia] and residence in a home classified in the highest wire-code category
according to either wire-code classification. There were no positive or statistically
significant dose-response trends, and results were not materially changed when
adjusted for potentially confounding variables.”
It is not possible to show the absence of a cancer effect from ELF-EMF, since it
is not possible to prove a negative. Scientists can quantify the size of an effect
if it exists, or they can set limits on the magnitude of a possible effect, but they
1
Risk by age 21 of selected childhood disorders (cases per 100,000): acute lymphocytic
leukemia (60), cerebral palsy (200), type I diabetes (200), autism (300), central nervous system
birth defects (300), schizophrenia (300), hypospadias-males (400), severe mental retardation
(400), congenital heart defects (600), asthma (6000) (Kaiser, 2003).
233
P1: GFZ
234 9. Electromagnetic Fields and Epidemiology
Figure 9.1. Childhood cancer and electricity power consumption. The incidence per million
children of all childhood cancer and childhood leukemia in Denmark are plotted from 1945
to 1990. On a different scale is plotted the per capita use of electrical power in 1000 GWh per
1 million inhabitants (Adair, 2000).
cannot state that the effect does not exist. Figure 9.1 displays childhood cancer and
leukemia rates as a function of time. Alongside the cancer data is a graph of the
use of electrical power in the United States as a function of time. The lack of an
association between cancer, which stayed relatively constant, and gross electrical
use, which grew tremendously, does not prove that there is not a link between EMFs
and cancer. But this graph shows that there is no significant link between cancer and
power lines.
In the following subsections we calculate direct EMF strengths from power lines,
E-fields in the body from power lines, induced E-fields in the body from the time-
varying magnetic fields of power lines, and E-fields in the body from naturally
occurring thermal fluctuations. These calculations will show that thermal fluctua-
tions in the body cause the largest E-fields in the body.
9.1.1 EMFs from Power Lines

Only 1 mW (10−12 of transmitted power) is radiated from a 10-km section of a 60-Hz,
500-MW power line. The reason for this radiation to be so low is that a 60-Hz wave
has a long wavelength of 5000 km. (We live in the near-field region.) The electric field
(E-field) from a power line is determined from its charge distribution (or voltage)
using Gauss’s law. The magnetic field (B-field) is determined from current flows
using Ampere’s law. B-fields are of possible concern since they easily penetrate the
human body to induce internal electric fields. On the other hand, external E-fields
are of less concern since the highly conducting human body shields most of the
external E-field. State regulations usually limit fields from transmission lines to
10 kV/m for E-fields and 200 milliGauss (mG) = 20 microTesla (μT) for B-fields.
(The milliGauss unit is used in most US regulations, even though the SI unit for
B-fields is the Tesla, with 1 μT = 10 mG.) Some regulations seek to constrain power
P1: GFZ
9.1. Power Line Health Effects? 235
line B-fields to 2 mG. The current public standard for all locations is 1000 mG, since
pacemakers experience abnormal pacing characteristics above this threshold.
Americans resides in Earth’s magnetic field of 450 mG (45 μT) and a 60-Hz
background level that ranges from 0.5 mG to 4 mG with an average of 0.9 mG
(primarily from appliances, not from power lines). Five percent of homes have
fields above 2.9 mG and 1% are above 6.6 mG. Instead of calculating magnetic
fields from a typical three-phase, six-wire configuration, our example power line
will be a two-wire 500-MW, 500-kV transmission line that draws a current of I = 500
A from opposing directions. Ampere’s law applied to one 500-A wire at r = 30 m
gives a magnetic field
B = μo I /2πr = (4π × 10−7 T-m/A)(500 A)/(2π )(30 m) = 3.3 μT = 33 mG. (9.1)
Two wires with opposing currents, separated horizontally by r = 4 m, give a
dipolar field in the plane of the wires at a distance r = 30 of
B = (μo I /2πr )(r/r ) = 33 mG(4 m/30 m) = 4.4 mG. (9.2)
This can be reduced to 1.1 mG by reducing the dipole moment per unit length.
This is done by moving the wires closer together to r = 1 m, or by doubling the
observation distance r . Underground power lines have lower B-fields than above
ground lines because their current conductors are much closer together. However,
the B-fields can be larger when you walk over them because you are closer to them.
Improving wire configurations to reduce fields from existing power lines is very
expensive, while incorporating better designs into new power lines is moderately
expensive.
Motors and appliances produce dipole or quadrupole fields that diminish rapidly
with distance. Average fields at 30-cm distance are as follows: color television
(7 mG), microwave (4 mG), analog clocks (15 mG), electric razors (20 mG, 100 mG
at 15 cm), and hair driers (1 mG, 300 mG at 15 cm).
9.1.2 E-fields in Biomatter from Power Lines

The E-field produced by a power line inside the body (E int ) is determined from Am-
pere’s law, which requires the sum of real current and displacement current (ε∂ E/∂t)
be continuous (in the normal direction) at the body’s surface between air and
biomatter. The equality of current density, external and internal to the body’s sur-
face is
ε0 ∂ E ext /∂t = jε0 ωE ext = σ E nt + ε∂ E int /∂t = σ E int + jεωE int , (9.3)
where ε is permittivity and σ is the body’s electrolyte conductivity. Since the E-Field
inside the body (E int ) is much less than the E-field external to the body (E ext ), the
displacement current inside the body can be ignored (εωE int = 0). The ratio of
E-field inside the body to that outside the body is
E int /E ext = 3ε0 ω/σ = 3(9 × 10−12 coul/V m)(2π )(60 Hz)/(0.5/ohm m) = 10−8 ,
(9.4)
P1: GFZ
where σ = 0.5/ohm m for the body’s electrolyte and the factor 3 comes from spher-
ical boundary conditions. For certain situations, such as the field experienced by
the ankles of a barefoot person on wet pavement, the ratio can rise to 10−7 . The
range of external E-fields most of us experience varies between E ext = 10 kV/m at
the edge of right-of-ways of large power lines to the E ext = 100-V/m static field
above Earth’s surface. As these E-fields enter the body, they are reduced by a factor
of 10−7 to 10−8 to E int = 1–100 μV/m, with the most likely being 1–10 μV/m.
9.1.3 E-fields in Biomatter from Oscillating B-fields

Changing B-fields create E-fields inside the body from Faraday’s law (See Fig. 9.2).
This gives a line integral of an induced E-field around a loop from the rate of change
of magnetic flux inside the loop:

E · dl = ∂ B/∂t d(area). (9.5)
For an AC power line with a circular path in biomatter, the internal E-field is
E int = (ωr B0 /2) sin(ωt). (9.6)
The power line magnetic field in the body is essentially the same as the B-field (B0 )
directly outside the body. Using the proposed regulation of B0 = 2 mG = 0.2 μT
over a chest-radius of 0.1 m at ω = 2π × 60 Hz gives a maximum E-field of
E int = (2π )(60)(0.1 m)(0.2 μT)/2 = 4 μV/m. (9.7)
Figure 9.2. Electric fields in biomatter. Curve (a), the line marked “kB T” is the internal E-
field, induced by Johnson–Nyquist thermal fluctuations of ions in biomatter. Section 9.1
calculates 20 mV/m at 60 Hz. Curve (b), the line marked “0.1 mT” is the induced E-field
for a large power line field of B = 0.1 milliTesla = 100 μ T = 1000 mGauss. The value at
60 Hz agrees with the Section 9.1 value of 4 μV/m for a power line field 0.2 μT (2 mG). The
curve (a) field is 10,000 greater than that given by curve (b) for power lines at the proposed
regulation level of 0.2 μT (2 mG) at 60 Hz. Curves (c), physiological responses of internal
fields, are considerably above curve (b) (Adair, 2000).
P1: GFZ
9.1. Power Line Health Effects? 237
Now we consider the different experience of walking in a constant magnetic field

along the equator in Earth’s B-field. The walk gives a transient current, resulting
in an internal E-field, which is balanced by polarization charge,
E int = v × B = (0.1 m/s)(45μT) = 4.5 μV/m (9.8)
for a slow walk of v = 0.1 m/s (0.4 km/hr). A fast run of 10 m/s gives E int =
450 μV/m, which is the same as the induced internal E-field from time-varying
B-fields at the edge of a 200-mG right-of-away.
Induced E-fields and current flow can also be obtained by rotating the body east
or west in Earth’s B-field. A 45◦ (π /4 rad) tilt of the head in the mid latitudes
(equator at 300 mG, poles at 600 mG) done slowly over 2 s creates the same internal
field as that induced by 2-mG power line fields:
E int = v × B = (r ω)(B) = (0.2 m)(π/8 − s)(45 μT) = 4 μV/m. (9.9)
E int = 50 μV/m for a fast nod and E int = 200 μV/m for a high diver doing flips.
9.1.4 E-fields from Internal Thermal Fluctuations

The E-field of a proton at a distance r = 10 Bohr radii (1 Bohr radius = 0.05 nm) is
huge:
E = (1/4πε0 )e/r 2 = (9 × 109 V m/C)(1.6 × 10−19 C)/(10 × 0.05 nm)2
= 6 × 109 V/m, (9.10)
where e is the charge of an electron and ε 0 is the permittivity of free space. It is not
surprising that Brownian ion motion in the body gives E-fields that exceed exter-
nally caused E-fields. The Johnson–Nyquist root mean square (rms) noise voltage is
Vrms = (4RkB T f )1/2 , (9.11)
where R is the sample resistance between measurement points and f is the
frequency interval for the measurement. For a cube of edge length d, the resistance
R = ρ/d = 1/σ d, where ρ is resistivity and σ is conductivity. The E-field from the
noise voltage has an rms magnitude of
E rms = Vrms /d = (4kB T f /d 3 σ )1/2 . (9.12)
We consider here the smallest biological cells, which have a diameter of about
20 μ, since they experience the largest Erms fields. (This is consistent with the fact
that the human body contains about 100 trillion cells.) The electrolyte between the
cells has a conductivity of about σ = 0.5/ohm m. Using kB T = 1/40 eV at room
temperature and f = 100 Hz, which more than surrounds the 50–60 Hz region,
gives
E rms = [4(1/40 eV)(1.6 × 10−19 J/eV)(100 Hz)/(2 × 10−5 m)3 (0.5/ohm m)]1/2
= 20 mV/m (9.13)
in the electrolyte surrounding 20-μ cells. Note that E-fields generated by thermal
fluctuations in biomatter of 20 mV/m are 10,000 times larger than the 4-μV/m
P1: GFZ
E-fields induced by 2-mG power line fields. This is one of the reasons why physi-
cists remain skeptical about potential health effects of power lines at the 2-mG
level. The 1996 APS Panel on Public Affairs examined biomedical epidemiology
issues beyond pure physics, but the above results gave impetus to study the issue.
Some have called for “prudent avoidance” of calling for regulations in cases where
there might be a health issue. The APS group concluded that it was imprudent to use
prudent avoidance to establish such public policy. At some point the risk has to be
compared to the mitigation cost to save a life. We used such quantitative approachs
in Section 7.4 when examining the issue of nonnuclear accidents with nuclear war-
heads. This approach values a life at about $5 million, but EPA recently used the
concepts of quality-adjusted years and willingness to pay for mitigating measures to
obtain more detailed analysis of the question risk. Clearly an average young life
is worth more than an average older life when quantified with respect to income
producing capacity.
9.2 Epidemiology
Epidemiology examines disease and health in human populations by identifying
associations between occurrence of diseases and hypothesized causes. Through
statistical methods, epidemiologists search for correlation between a disease and en-
vironmental or other factors. However, correlation does not prove causality. Other
criteria must be met before an association may be considered certain, likely, prob-
able, possible, weak, or not all.
British epidemiologist Austin Bradford Hill suggested (1961) the now well-
known criteria to judge whether a statistical association gives a specified de-
gree of confidence in the crediting of causality to a link between exposure and
disease. In other words, what criteria should be addressed beyond apparent asso-
ciation when a researcher is considering the relevance of epidemiology studies.
Hill’s criteria have been adapted below to determine whether ELF-EMF cause
cancer.
1. Strength: Is there a strong correlation between disease and ELF-EMF fields?

2. Consistency: Have the same results been obtained by other research in other
locations?
3. Specificity: Do ELF-EMFs produce the same types of cancer in similar proportions
to other groups similarly exposed?
4. Temporality: Since there is a latency period for cancer, are the ELF-EMF diseases
in the present the same as in the past?
5. Biological Gradient: Do higher doses of ELF-EMF cause more cancer than lower
doses? Is there an approximate proportionality of risk and dose, as in the case
of the probability of additional lung cancer (incidence and mortality) and the
number of cigarettes smoked per day?
6. Plausibility: Do the biological data on conjectured cancer promotion by ELF-
EMF converge on a plausible, consistent biological-biophysical mechanism? A
P1: GFZ
9.2. Epidemiology 239
mechanism in this case is a discernible physical path (on a cellular level) at-
tributable to ELF-EMF in the body.
7. Coherence: One should expect coherence between the data and the mechanism. In
general, most mechanisms that attempt to connect ELF-EMFs and cancer would
predict that enhanced exposures to ELF-EMFs enhance cancer rates.
8. Experiment: Are the various in-vitro (cells in culture) and in-vivo (complete liv-
ing systems) ELF-EMF experiments consistent among themselves and with a
theoretical mechanism?2
9. Analogy: Is the connection between ELF-EMF and cancer analogous to situations
where the proof is more substantial? Does one have to have new biophysics to
understand this connection?
American Physical Society (1996) and National Academy of Sciences (1997) pan-
els concluded that, based on Hill’s criteria, a link between ELF-EMFs and cancer
has not been determined. We now use a statistical procedure to interpret data from
a hypothetical case involving an unnamed suspected carcinogen. This procedure
allows us to determine the likelihood of whether a correlation between disease and
exposure is random oractual.
9.2.1 Relative Risk

The relative risk RR of contracting a disease is given by
RR = disease rate among exposed population/disease rate among
nonexposed population. (9.14)
When RR is large, such as the case of RR = 20 of contracting lung cancer from
smoking one pack of cigarettes a day, it is easy to find biomedical data to satisfy
Hill’s criteria. However, when RR is small, say, less than 2 for a rare disease, associ-
ation between exposure and contraction of a disease is often rejected, unless other
data are available. The statistical significance of the association must be examined.
Very high RR may well signal a strong association.3 Yet any study that seeks to
determine causality between a suspected carcinogen and a disease must correct
for systematic errors caused by biased data. Another problem arises when choice of
populations being studied results in data that mask association. For example, one
may compare exposure/disease of a working population to exposure/disease in
2
Robert Koch’s postulates (1890) to determine whether a particular microbe causes a par-
ticular disease are as follows:
1. The microbe under suspicion must be found in all cases of the disease.
2. It must be isolated and grown in the laboratory.
3. It must be injected into a healthy animal or person to see if it causes systems of the disease.
4. Microbes isolated from this new case must be identical to the original microbe.
3
By this definition the RR of severe acute respiratory syndrome (SARS) was infinity since it
did not previously exist. For an analysis, see S. Riley, C. Fraser and C. Donnelly, Transmission
dynamics of the etiological agent of SARS in Hong Kong, Science 300, 1961–1966, 2003.
P1: GFZ
the general population. But a masking factor may be that the working population
experiences better health than the nonworking population. Thus, the comparison of
the effect of a carcinogen is more accurately determined if the comparison is made
between workers that were exposed and workers that were not exposed. More
data gives better statistics. In Section 7.1.5 we determined sample sizes needed to
determine if cancer could be shown to be correlated with exposure to low doses of
radiation from medical x-rays.
Epidemiology must also remove effects of a confounding bias, in which another
variable is related to both the possible carcinogen and the disease. For example,
lung cancer would be expected to be associated with a lifestyle involving both alco-
hol consumption and smoking. If only alcohol consumption as it relates to cancer
is studied, one might conclude that alcohol causes lung cancer. Yet we know that
smoking directly causes cancer, while heavy drinking may just degrade the general
health of the body. A better analysis would be one that compared incidents of lung
cancer with the intensity of smoking among nondrinkers, or at least among those
with similar drinking habits. A more complex confounding situation is the case of
uranium miners that smoke. Miners who spend time in a radioactive environment
and who also smoke have enhanced cancer rates over nonsmoking miners in the
same environment. Radioactivity and smoking acting synergistically to give a can-
cer rate that is greater than the sum of the individual cancer rates for smoking or
radioactivity. Thus, good epidemiology uses cohort groups of subjects having similar
traits, such as similar age, gender, race, income, and so forth. Iceland’s population
is often examined because of its great homogeneity.
A case-control study identifies persons who have a disease (the cases) and those
who do not (the controls). For simplicity, we assume that all of the individuals
exposed to a suspected carcinogen had the same amount of exposure. The results
of case-control study such as this can be displayed in a 2 × 2 matrix, as we have
done in Table 9.1, in which each cell contains the number of persons in the four
possible situations: a = diseased with exposure, b = diseased without expose, c =
not diseased with exposure, and d = not diseased without exposure.
As an example, consider the data in Table 9.2 from a hypothetical study of 200
individuals: 100 were exposed to a carcinogen and 20 became diseased; 100 were
not exposed, but 10 became diseased:
The relative risk of getting the disease is
RR = rate among exposed/rate among the unexposed

= (20/100)/(10/100) = 2.0 (9.15)
Table 9.1. Case-control study

Cases Controls Total
Exposed a c a+c
Nonexposed b d b+d
total a+b c+d a+b+c+d
P1: GFZ
9.2. Epidemiology 241
Table 9.2. Hypothetical epidemiology case-control study

Exposed 20 80 100
Nonexposed 10 90 100
total 30 170 200
Is this outcome statistically significant? Does a true association exist between

disease and exposure? Or, could the apparent association be a chance occurrence
from data fluctuations? The first step in answering this is to calculate the expected
values of the four parameters (a , b, c, d) on the basis that no association exists between
exposure and disease. In our analysis, expected values of each group are compared
to actual values to determine the statistical significance of an association between
disease and exposure. Under the assumption of no connection between disease and
exposure, the expected value E j of all four parameters is based on the ratio of the
total number diseased (a + b) to the total population (a + b + c + d). This should
be the same ratio for each cell that is given if there were no disease. For example,
the expected value E a for the number of diseased and exposed persons (a ) without
a viable disease divided by the total number exposed (a + c) is equal to the ratio
above, giving
E a /(a + c) = (a + b)/(a + b + c + d). (9.16)
This gives the expected value of a ,
E a = (a + c)(a + b)/(a + b + c + d) = (20 + 80)(20 + 10)/(20 + 10 + 80 + 90) = 15

(9.17)
A generalized form for all the expected values is obtained from this,
E = (row total)(column total)/grand total. (9.18)
Applying this formula to each cell in Table 9.2 gives Table 9.3, the 2 × data matrix
of expected values (with hypothetical data in brackets).
Table 9.3. Expected values for no association between

exposure and disease
Exposed 15 85 100
[20] [80]
Nonexposed 15 85 100
[10] [90]
Total 30 170 200
Hypothetical data in brackets.

P1: GFZ
9.2.2 χ 2 Without an Association Between Cause and Disease

The value of chi-squared (χ 2 ) is a measure of difference between expected values
E of the parameters with no association and the observed data O, which may carry a
causal connection to disease. This approach gives statistical confidence in assigning a
degree of association between exposure and disease. The χ 2 value is determined
by summing over the four cells in the 2 × 2 table, that is
4
χ2 = j=1
(|0j − E j | − 0.5)2 /E j . (9.19)
The value 0.5 is subtracted from absolute value of |O – E| in each cell to make the
naturally discrete data continuous (since one cannot be partially diseased).
9.2.3 Observed Significance Level, the p-value

From the χ 2 value, one can determine the observed significance level, called the p-value,
which is the probability that an apparent association between a disease and a suspected
carcinogen is caused by randomness. If the p-value is very small, the hypothesis of no
correlation between cases and disease should be rejected. For our example, we obtain
χ 2 = 3.18. Large values of χ 2 result from large differences between expected val-
ues E (no association) and actual data O, giving a strong association of exposure
with disease with only a slight chance of there being a false conclusion caused
by randomness of data. Using standard statistical tables of χ 2 versus probability
p, the value χ 2 = 3.18 gives a p-value = 0.08, which means there is only an 8%
chance that the association is caused by a statistical fluctuation. Since 8% is 1 part
in 12, these data may appear conclusive that there is a true causal relationship be-
tween exposure and disease. However, risk analysis can be vulnerable to bias and
confounding effects. In our case more epidemiology data are needed, along with
biomedical analysis to determine the causal connection. Hill’s criteria should be
considered before promoting public policy on the basis of these results. The lit-
erature contains many papers that claim weak associations, but upon reflection
their conclusions are rejected. Epidemiology has limited use for data reflecting low
relative risks for rare diseases.
9.2.4 Upper and Lower Limits

An alternative approach to p-values is to determine degree of confidence by using
standard deviations σ of the mean. Because epidemiology is often uncertain and
because it strongly affects public policy, the +/− error bars are wisely chosen at a
wide 1.96σ , covering 95% of events. This is a higher standard than the 1σ error bars
usually used in scientific measurements, covering 68% of events. The variance σ 2 of
a sample of N measurements of the same type is the variance of one measurement
divided by the number of measurements, or σ 2 /N. The variance of each of the four
cells is the inverse of number of individuals represented in the cell. Since these
four variances are statistically independent, the total variance σ 2 is the sum of the
P1: GFZ
Problems 243
individual variances:
σ 2 = 1/a + 1/b + 1/c + 1/d. (9.20)
The data set gives
σ 2 = 1/20 + 1/10 + 1/80 + 1/90 = 0.173, (9.21)
which gives σ = 0.417. The natural logarithm gives the lower limit LL and the
upper limit UL of the confidence interval. For the case of a 95% interval (+/−
1.96σ ) centered about RR = 2,
ln(LL) = ln(RR) − 1.96σ = 0.693 − 0.817 = −0.124 (9.22)

ln(UL) = ln(RR) + 1.96σ = 0.693 + 0.817 = +1.510. (9.23)
These results give the upper and lower limits,
LL = e −0.124 = 0.88 and UL = e 1.510 = 4.5. (9.24)
9.2.5 Limits Versus p-values

The relative risk RR = 2.0 is bounded between limits of 0.88 to 4.5 for 95% confi-
dence. The intervals are not symmetric since the probability curve is not symmetric
about RR = 2 and there are no negative RR values. However, note that in spite of
this asymmetry, the ratios RR/LL = 2.3 = UL/RR. The unequal spacing makes
association appear stronger than it really is. The LL of 0.88 appears to be very close
to 1, while the UL of 4.5 seems fairly dangerous. Because of the uncertainties in
epidemiology, we prefer to quote a range of values that use the wider 2σ , 95% con-
fidence limits (mean value 2.0, bounded from 0.88 to 4.5). The limits are preferable
to quoting a relative risk with a p-value (2.0, 7.5%), which implies a robust link
between pollution and cancer from very minimal data. The key to understanding
these data is to know whether they contain hidden biases, such as preselection of
cancer cases in a cluster, or confounding effects that blur the result.
Problems
9.1 Regulation of 2 mG. A 500-A, single-phase power line has two lines separated
by 2 m. What is the distance from the line for a 2-mG (0.2 μT) proposed
limit?
9.2 Ankles = 10 stomachs. Power line-induced, internal electric fields near a per-
son’s ankles on wet pavement can be a factor of 10 greater than the fields
induced near the stomach. Why is this so?
9.3 E-fluctuation. Time-varying B-fields produce internal electric fields in biomat-
ter measuring 1–10 μV/m. What distance from a hydrogen ion gives the same
E-field? How much is this reduced by dielectric screening of water?
P1: GFZ
9.4 Walk versus nod. How fast a 90◦ head rotation to east or north at the equator
is equivalent to a 5 m/s walk to the east? What is physically different between
walking and nodding for E-fields?
9.5 Magnetosomes on the brain. The most likely EMF link to health could
be through very dilute magnetite particles in the brain. (a) Write a differ-
ential equation describing the rotation of a spherical, single-crystal mag-
netite particle that takes into account inertia, viscous drag, Earth’s magnetic
field, and an external oscillating magnetic field, but ignoring thermal colli-
sions. (b) Using symbols show that small magnetosomes experience small
torques from oscillating magnetic fields and large magnetosomes experi-
ence large viscous torques. Is there a size that maximizes magnetosome
oscillations?
9.6 B2 dependence. Why might ELF-EMF health effects, if they exist, be dependent
on the square of the B-field strength? If this is true, how might this dependence
be discovered with epidemiology?
9.7 Cyclotron radius. (a) What is the cyclotron frequency for calcium ions
(Ca+ ) in Earth’s magnetic field of 45 μT? (b) What is the radius of a cal-
cium ion at room temperature energy of 0.02 eV? At 3 eV to break DNA
bonds? (c) What is the approximate collision frequency of calcium ions in
water?
9.8 Cell phones. A cell phone consumes 10 W, with 10% appearing as radiation at
a frequency of 900 MHz. (a) What is the power flux to the head if the cell phone
antenna is 5 cm away? (b) How does this compare to the body’s blackbody
radiation flux from 35◦ C skin?
9.9 Cell-phone standards. If the 900-MHz signal of problem 9.8 penetrates 2 cm
into the body, what is the dose in watts/kg? How does this compare with the
standard of 1 W/kg needed to confine temperature rise to 0.1◦ C? What do you
estimate for the temperature rise?
9.10 Healing magnets. Some sick persons attach magnets to wounds in a quest
for healing. (a) Does a healing magnet stick to the icebox door when there
are five sheets of paper between magnet and door? What does this say about
magnetic penetration? (b) Determine how are healing magnets constructed
and whether this makes sense? (c) How can magnetic field gradients induce
currents and give microforces on moving blood?
9.11 Epidemiology. A group of 100,000 persons is exposed to a pollutant and 20,000
develop a disease. Of an unexposed group of 100,000 persons, 15,000 develop
the disease. (a) Based on these data, what is the relative risk of developing
the disease from exposure to the pollutant? (b) What are the expected values
if there is no link between the pollutant and disease? (c) What are χ 2 , p, RR,
UL, and LL, and how do they compare with the text example?
9.12 Hormone replacement therapy. The Women’s Health Institute reported in
2003 that of 8506 women on hormones, 199 developed invasive breast cancer,
compared with 150 cases for the 8102 women who took placebos. What are
χ 2 , p, RR, UL, and LL for these data?
P1: GFZ
Bibliography 245
Bibliography
Adair, R. (1991). Constraints on biological effects of weak extremely low-frequency electro-
magnetic fields, Phys. Rev. A 43, 1039–1049.
———(2000). Static and low-frequency magnetic field effects: Health risks and therapies,
Reports Prog. Phys. 63, 415–454.
Bennett, W. (1994). Cancer and power lines, Phys. Today 47(4), 23–29.
Bennett, W. (1994). Health and Low-Frequency Electromagnetic Fields, Yale University Press,
New Haven, CT.
Brouder, P. (1995). The Great Power-Line Cover-Up: How the Utilities and the Government Are
Trying to Hide the Cancer Hazard Posed by Electromagnetic Fields, Little Brown, Boston, MA.
Foster, K., D. Bernstein and P. Huber (Eds.) (1993). Phantom Risk: Scientific Inference and the
Law, MIT Press, Cambridge, MA.
Hafemeister, D. (1966). AAPT Resource Letter: biological effects of low frequency electro-
magnetic fields, Am. J. Phys. 64, 974–981.
Hafemeister, D. (Ed.) (1999). Biological Effects of Low-Frequency Electromagnetic Fields, Ameri-
can Association of Physics Teachers, College Park, MD.
Kaiser, J. (2003). Science 301, 162–163.
National Research Council (1966). Possible Health Effects of Exposure to Residential Electric and
Magnetic Fields, National Academy Press, Washington, DC.
Polk, C. and E. Postow (Eds.) (1996). CRC Handbook of Biological Effects Electromagnetic Fields,
CRC Press, Boca Raton, FL.
Rothman, K. (1986). Modern Epidemiology, Little Brown, Boston, MA.
Samuels, M. and J. Witmer (1999). Statistic for the Life Sciences, Prentice Hall, Upper Saddle
River, NJ.
Slovic, P. (1987). Perception in Risk, Science 236, 280–285.
Wilson, R. and E. Crouch (2001). Risk-Benefit Analysis, Harvard University Press, Cambridge,
MA.
P1: GFZ
Energy
10 The Energy Situation ..................................................................... 249

10.1 Introduction ........................................................................... 249
10.2 Energy Orders-of-Magnitude .................................................... 260
10.3 Fossil Fuel Models ................................................................... 262
10.4 Energy Rates of Change............................................................ 268
10.5 Population and Sustainability.................................................... 269
10.6 Single and Combined Cycle Power Plants ................................... 271
10.7 LNG Explosions ...................................................................... 274
11 Energy in Buildings ....................................................................... 279

11.1 Heat Transfer .......................................................................... 280
11.2 Single- and Double-Glazed Windows ......................................... 283
11.3 Annual Heat Loss.................................................................... 284
11.4 Energy Standards .................................................................... 288
11.5 Scaling Laws for Buildings........................................................ 290
12 Solar Buildings.............................................................................. 299

12.1 Solar Flux............................................................................... 299
12.2 Solar Collectors ....................................................................... 302
12.3 Integrated Solar Flux................................................................ 304
12.4 Solar Hot Water....................................................................... 307
12.5 Active Solar Space Heat............................................................ 309
12.6 Passive Solar Thermal Flywheel................................................. 310
13 Renewable Energy ......................................................................... 316

13.1 Sustainable Energy .................................................................. 316
13.2 Photovoltaic Solar Power.......................................................... 317
13.3 Solar Thermal Power................................................................ 320
13.4 Hydropower........................................................................... 321
13.5 OTEC and Thermoclines .......................................................... 324
13.6 Wind Power ........................................................................... 326
247
P1: GFZ
13.7 Tidal and Wave Power ............................................................ 327

13.8 Geothermal Power ................................................................. 330
13.9 Biomass Power ...................................................................... 332
13.10 Fusion Power ....................................................................... 333
14 Enhanced End-Use Efficiency .......................................................... 343

14.1 Heat/Cold Storage in Large Buildings ...................................... 344
14.2 Improved Lighting................................................................. 347
14.3 Improved Windows ............................................................... 351
14.4 Heat Pumps .......................................................................... 353
14.5 Improved Appliances ............................................................. 356
14.6 House Doctors....................................................................... 359
14.7 Cogeneration ........................................................................ 364
14.8 Utility Load Management ....................................................... 365
14.9 Energy Storage ..................................................................... 371
15 Transportation............................................................................... 378
15.1 Automobile Energy Basics....................................................... 380
15.2 Corporate Average Fuel Economy (CAFE) ................................. 384
15.3 IC Engines ............................................................................ 386
15.4 Hybrid Cars .......................................................................... 388
15.5 Hydrogen Fuel-Cell Car.......................................................... 390
15.6 Safety................................................................................... 394
15.7 Transportation Potpourri......................................................... 396
16 Energy Economics.......................................................................... 402

16.1 Basic Economics .................................................................... 402
16.2 Discounted Benefits and Payback Periods.................................. 409
16.3 Cost of Conserved Energy ....................................................... 412
16.4 Minimum Life-Cycle Costs ...................................................... 414
16.5 Energy Tax Credits ................................................................. 415
16.6 Petroleum Economy ............................................................... 417
16.7 Imported Oil, Synfuels, and Gasohol......................................... 419
16.8 Plutonium Economy............................................................... 422
248
P1: GFZ
10
The Energy Situation
Heat is work and work is heat,

That’s the First Law of Thermodynamics
The Second Law of Thermodynamics,
Heat cannot go by itself from one body to a hotter body,
The Universe is going to cool down,
Then there will be no more work,
That’s Entropy Man.
[M. Flanders and D. Swann, The Laws
of Thermodynamics, Angel Records]
10.1 Introduction
The world continues to indulge its dangerous addiction to Middle East oil. Seem-
ingly little has been learned from the oil embargo of 1973–74 that resulted in long
gas lines, double-digit inflation and consumer helplessness. Since then the effects
enhanced end-use efficiency (better technologies) and increased global petroleum
production dropped gas prices to pre-embargo values (corrected for inflation), but
then rose to $3.50/gal by 2006 and $78 per barrel for crude oil. But petroleum
supplies continue to threaten national security. Oil was the major factor for US
participation in the 1991 Gulf War. Oil was a continuing factor in Iraq for the in-
vasion of 2003. The United States is projected to need 28 million barrels per day
(1 Mbbl/day = 2.1 quads/year = 2.1 × 1015 BTU/year) by 2025, compared to 20
Mbbl/day in 2003. Yet the United States will be able to produce only one-third
of this amount. The petroleum situation would be even more bleak if it were
not for improved automobile fuel economy, which had doubled by 1985, but is
now back-sliding. Until alternate energy becomes competitive, natural gas and
coal will be the bridging fuels that will supply portable energy as petroleum be-
comes expensive and less plentiful. And the burning of fossil fuels will proba-
bly double CO2 levels during the next century, causing an unknown amount of
249
P1: GFZ
250 10. The Energy Situation
Figure 10.1. US energy consumption by fuel, 1980–2030. Data in quadrillion Btu/year

(1 quad/year = 1015 Btu/year). Fossil fuel use is projected to grow considerably in the
next quarter century, especially petroleum and natural gas, but also coal. Renewable en-
ergy is projected to grow only slightly and nuclear energy is projected to decline, es-
pecially after 2030. [Annual Energy Outlook, Energy Information Administration (EIA),
2006]
climate change. Meanwhile, fossil burning continues to plague urban air qual-
ity. Natural gas was expected to be the bridging fuel of the future, but dur-
ing 2003–05 its future supplies from Canada became less certain and its price
greatly increased. The electrical utilities responded mightily by increasing plans
for coal plants from 2 to 100, and for nuclear plants from 0 to 10. See Figs. 10.1
and 10.2.
Prior to the 1973–74 oil embargo, fossil fuels supplied 92% of US energy needs.
Today they continue to dominate, comprising 86% of US energy use in 2004. Energy
Figure 10.2. Global energy

consumption. Projections for
global energy follow the same
trends as US projections in
Fig. 10.1, but they are a factor
of 4 larger. [International Energy
Outlook, EIA, 2006]
P1: GFZ
Table 10.1. US energy consumption by sector, 2004.

Sector 2004 energy energy (%) 2001 cost cost (%)
Residences 21.0 quads 21% $167 billion 24
Commercial bldg. 17.4 quads 17% $127 billion 18
Buildings (total) 38.4 quads 39% $294 billion 42
Industry 33.2 quads 33% $135 billion 19
Transportation 28.0 quads 28% $270 billion 39
Total 99.7 quads 100% $699 billion = 7% of GDP
Electricity generation consumed 40 quads, which was mainly used in buildings

and industry (Energy Information Administration, 2006).
consumption rose 23 quads1 during the 30-year period.2 The medium gainer was
petroleum, which grew 5 quads (from 35 to 40), while the small gainers were natural
gas, which stayed about constant at 23 quads and renewable energy which grew
about 1 quad in a normal hydro year (from 5 to 6). The large gainers were nuclear
power, which grew 7 quads (from 1 to 8) and coal, which grew 10 quads (from 13
to 23). The per capita total cost of energy is about $2500 ($700 billion/290 million
population) and includes the broader cost to society as a whole from industry,
commerce, and government. The cost of energy to a typical US household in 2001
was $2911 with $1392 for homes ($915 for electricity, $386 for natural gas, $90 for
fuel oil, $99 other) and $1519 for gasoline. See Table 10.1.
Before the oil embargo, US energy use was growing 4.4% per year, with a dou-
bling time of 15 years. Electricity was growing even faster at 7% per year. The
Atomic Energy Commission projected in 1972 that the United States would more
than double its energy use to 160 quads/year (160 × 1015 Btu/year) by 2000. In
1974, the AEC lowered this to 140 quads, only for the new Energy Research and
Development Agency (ERDA) to further lower it in 1976 to 124 quads. The reduced
projections were still too large as the United States used 99.3 quads from all en-
ergy sources in 2000, which was the same result as the zero growth projection of the
then-controversial 1974 Ford Foundation Energy Policy Project. Amory Lovins in
his 1976 paper, “Energy Strategy: The Road Not Taken,” got the right answer, but
for the wrong reasons, since he over-promoted sun, wind and biomass, which did
not have a great impact. But Lovins’s analysis also championed enhanced end-use
efficiency, which accounted for the major difference in US energy between 1975
and 2000. See Figs. 10.3 and 10.4 and Table 10.2.
The American Physical Society’s Panel on Public Affairs (APS-POPA) 1996 study,
Energy, the Forgotten Crisis, chaired by this author, concluded that serious, long term
shortages of cheap energy were being largely ignored. Such a lack of concern for
1
20 quads = 20 × 1015 Btu = 3.5 Gbbl.
2
Data were updated during July 2006, using figures from the Energy Information Admin-
istration and other sources.
P1: GFZ
252
SVNY342-Hafemeister
March 24, 2007
10. The Energy Situation

11:21
Figure 10.3. US energy flow in 2004 (quad/year). This rendition displays energy production by source on the left, flowing into energy
consumption by sector (including waste energy). Electricity generation (40 quads) is used mostly in buildings and industry. [Annual
Energy Review, EIA, 2006]
P1: GFZ
Figure 10.4. US electrical generation by fuel (billion kWh/year). Increased electricity gen-
eration is projected to come primarily from coal. Electricity consumption is projected to rise
by a factor of 4 over 60 years, from 1392 billion kWh in 1970 to 5619 billion kWh in 2030.
A 1-GWe plant, operating continuously produces 8.8 billion kWh/year. This translates to
average power of 161 GWe in 1970 and 646 GWe in 2030. Peak power is about twice these
numbers. [EIA, 2006]
future energy supplies may not be surprising as fuel supplies generally seem ample
and prices are relatively low. But to neglect energy difficulties is imprudent. To set
the stage for our discussions in Chapters 10 to 16, we follow with an updated
summary of the APS-POPA study. See the data in Figs. 10.5 to 10.13.
Adequate energy supplies are crucial (Chapter 16). Economic growth, coupled with
1% annual US population growth and a 1.3% global population growth, increase
energy demand (Chapter 16).
Coupling between energy and the economy has lessened but it still is a reality. Increased
energy efficiency and modal shifts decreased the ratio of US energy use to GDP by
35% (18,000 Btu/$ in 1973 to 11,800 Btu/$ in 2000, all in 1996 dollars). This decrease
was due, in part, to the fact that GDP rose 34% while energy use rose only 15%
from 1990 to 2000. Energy use per capita has been relatively constant, dropping by
14% after the oil embargo, but returning to the 1973 level of 3.4 × 108 Btu/person
(60 oil barrels/person/year) in 2003.
Table 10.2. US energy consumption by fuel before and after the oil embargo
Fossil energy Nuclear/renewable
oil ngas coal total nuc hyd biom other Total fossil (%)
1973 34.8 22.5 13.0 70.3 0.9 3.0 2.3 0.04 76.6 92
2000 38.5 24.1 22.7 85.3 8.0 2.8 2.9 0.3 99.3 86
2004 40.1 23.1 22.5 85.7 8.2 2.7 2.9 0.4 99.7 86
Consumption of oil, natural gas, and hydroelectric power changed relatively little from 1973 to 1995,
while coal grew 9.7 quads, nuclear power grew 7 quads, and renewable energy grew 1 quad/year.
(1 Mbbl/day = 0.365 Gbbl/year = 2.12 quads/year, EIA, 2006)
P1: GFZ
Figure 10.5. US energy-use

per capita and per dollar of
gross domestic product. Data
and projections are normal-
ized to 1 in 1980. Over the
1980–2030 period, energy use
per capita is projected to rise
5%, while energy used per
GDP dollar is projected to drop
60%. [EIA, 2006]
The world aspires to greater energy use. US energy use increased by 15% from 1990
to 2000, while energy use by the developing Asian countries, China, India, and
South Korea rose by 40%. Within several decades, the developing countries could
surpass OECD countries’ energy use.
Fossil fuels are the primary energy source. Fossil fuels provided 86% of US’s
98.2 quads from all sources in 2003, with 40% from petroleum, 23% from natu-
ral gas, and 23% from coal.
Limited US petroleum production. US proven oil reserves dropped from 32 bil-
lion barrels (1 gigabarrel = 109 bbl = 1 Gbbl) in 1977 to 23.1 Gbbl in 2004. The
remaining recoverable oil in the United States, both discovered and anticipated,
is perhaps 175 Gbbl (124.1 Gbbl unproved in 2004), which is much less than the
amount already produced (195 Gbbl by 2004). In spite of Alaskan production, US
production dropped from 9.6 Mbbl/day in 1970 to 5.4 Mbbl/day (plus 3.2 other
domestic production) in 2004, compared to net import of 12.1 Mbbl/day. Lower
48 states production dropped from 9.4 Mbbl/day in 1970 to 4.8 Mbbl/day, while
Figure 10.6. World energy

consumption by region
(quads/year). Energy use
by non–OECD nations is
projected to quadruple during
1980–2030, while energy use
in OECD nations is projected
to rise only 55%. [EIA, 2006]
P1: GFZ
Figure 10.7. Discovery decades for the 100 largest oil fields and 100 largest gas fields in
the US. These data are based on estimated ultimate recovery of the fields. Significant new
discoveries of large oil and gas fields are unlikely. [Largest oil and gas fields, EIA, 1993]
P1: GFZ
Figure 10.8. Domestic oil pro-

duction by source. Lower–48
onshore production is pro-
jected to drop by 50% from
its peak year of 1970 (Sec-
tion 10.3). Alaska peaked at
2 Mbbl/day in 1988, but is
one–half that by 2003. Produc-
tion from ANWAR could sus-
tain Alaskan production at 1
Mbbl/day. [EIA, 2006]
Figure 10.9. Petroleum con-

sumption, production and im-
ports, 1990–2030, (Mbbl/day).
US consumption is projected
to rise to 28 Mbbl/day in 2030,
with 18 Mbbl/day from im-
ports (60%). [EIA, 2006]
Figure 10.10. World crude oil

reserves by region, 1980–2006.
[EIA, 2006]
P1: GFZ
Figure 10.11. Natural gas proved

reserves by region, 1980–2006 Data
in trillion cubic feet (1 TCF = 1
quad). The predominant proved
reserves are in Russia (1680 TCF),
Iran (971 TCF), Qatar (911 TCF),
Saudi Arabia (241 TCF), United
Arab Emirates (214 TCF) and the
US (193 TCF), which have 69% of
the 6112 TCF total. [EIA, 2006]
Figure 10.12. Natural gas con-

sumption by sector, 1990–2030
(TCF/year). Electrical generation
with combined cycle gas turbines
was expected to raise natural gas
consumption considerably, but the
natural gas supplies have become
more uncertain. US consumption is
projected to rise from 22.4 TCF in
2004 to 26.9 TCF in 2030. [EIA, 2006]
Figure 10.13. US natural gas imports, 1990–2030 (TCF/year). Natural gas from Canada was
projected to be 5 TCF/year in 2025, but by 2005 this was lowered to 1.7 TCF. Imported
liquefied natural gas (LNG) was projected to rise from 0.5 TCF in 2003 to 2.1 TCF in 2025,
but in 2005 this was projection was raised to 4.3 TCF, increasing the number of LNG port
facilities needed. [EIA, 2006]
P1: GFZ
Alaskan production dropped from 2 Mbbl/day in 1988 to 1 Mbbl/day. The mean

estimate of production from the Arctic National Wildlife Refuge (ANWR) is that
it could provide 0.5 Mbbl/day, but this represents only 2% of US consumption.3
Discovery of new large US oil and gas fields seems unlikely as all fields discovered
since 1980 have been smaller than the top 100 previously discovered fields.
Reliance on oil imports creates problems. Net petroleum import in 2003 was
12.1 Mbbl/day in 2004. This cost $125 billion (at $75/bbl it is $330 billion), which
was 25% of the trade deficit and represents 8% of all imports. Net imports of crude
petroleum almost doubled since the oil embargo (6 Mbbl/day in 1973) to 60% of
present use. The OPEC fraction of global oil is projected to rise from 40% in 1990 to
57% in 2015. The global reserve was estimated in 2004 at 1300 Gbbl, with the largest
proved reserves reside in the Middle East (743 Gbbl), a region of violent instability,
followed by South America (103 Gbbl), Russia and East Europe (79 Gbbl), Africa
(102 Gbbl), and the Far East and Oceania (39 Gbbl). In North America, Mexico
(13 Gbbl) and the United States (21 Gbbl) were listed ahead of Canada until 2003,
when Energy Information Administration (EIA) raised its estimate of Canada’s
proved reserves from 6 to 179 Gbbl. This was in response to improved extraction
techniques in the Athabasca tar sands of Alberta, but the World Oil journal main-
tains the 6 Gbbl estimate since cost and environment are issues. The 10% petroleum
content and its depth below surface made this oil expensive, but new techniques
using multidirectional wells and steam injection improved the cost of extraction. But,
oil from tar sands is not competitive. US vulnerability on oil imports will ensure
US involvement in these regions. Europe and Japan have partially prepared them-
selves for future shortages in other ways, through high taxes on gasoline (bringing
costs to $5 per gallon) that raise revenue and moderate demand (Chapter 16).
Coal is the bridging fuel, while natural gas becomes uncertain. US proved natural gas
reserves dropped slightly from 200 trillion cubic feet (TCF) in 1983 to 193 TCF in
2004. (1 TCF provides one quad.) Natural gas consumption of 23.1 TCF in 2004 is
predicted to rise to 27 TCF by 2025. Total US technically recoverable resources might
be as high as 1300 TCF, but it is unclear how much of this can be converted to proved
reserves. In 2003, 15% of US natural gas came from Canada, but future growth in
supplies from Canada is uncertain. Natural gas supplies became more uncertain in
2003. Large coal reserves exist, but mining and burning coal entails environmental
problems.
Energy production and use adversely affect the environment. Routine energy use and
accidents release pollutants that harm air and water quality (Chapter 6), pre-empt
land and river use and adversely affect climate (Chapter 8).
Buildings and appliances consume 40% of US energy, presenting opportunities to reduce
energy requirements (Chapters 11, 12, 14). During the period 1973–1985, thermal
resistance of ceiling insulation in typical new US single-family houses increased
from R14.4 to R26.7, while wall insulation increased from R10 to R12.5. Houses
3
US Geological Service 1998 estimates of the technically recoverable ANWAR reserve in 1996
dollars: 3.2 Gbbl with a range of 0.7–7.0 Gbbl at $20/bbl and 5.6 Gbbl with a range of 2.3–9.5
at $25/bbl.
P1: GFZ
Table 10.3. US renewable energy production

hydro geoth wind solar biom. renewables energy total
1973 2.9 0.04 0 0 2.3 5.2 76.6 quads
2000 2.8 0.3 0.05 0.06 2.9 6.1 99.3 quads
2004 2.7 0.3 0.14 0.1 2.9 6.0 99.7 quads
2004 electric 31 1.6 1.6 0.1 6.8 41 451 avg. GWe
Energy is given in quads per year and average electrical from renewables in GWe (EIA, 2006).
built after 1988 consume only 59% as much natural gas as those built before 1980.
New refrigerators reduced energy consumption by 72%; energy use in refrigerators
dropped from 1800 kWh/year in 1974 to 500 kWh/year in 2001. Long-lived compact
fluorescent lamps use only 30% of the energy used by incandescent bulbs. Energy
use in new buildings can be decreased by an additional 50% with cost-effective
energy features.
The transportation sector petroleum use in the United States is critical as it accounts
for 65% of US petroleum and 27% of total energy (Chapter 15).4 The total number of
vehicles of all types doubled from 106 million in 1970 to 227 million in 2001 with
191 million licensed drivers. In 2001, US light duty vehicles traveled 2409 billion
vehicle miles, growing at 2.3% per year. Considerable fuel is saved through the
CAFE fuel economy standard of 27.5 miles/gal, which doubled the 1973 average
of 13.5 mpg. However, the light truck and SUV standard is only 20.7 mpg. In 2001
sales of light trucks and SUVs (20.9 mpg) equaled those of cars (28.6 mpg). This
reduced the average new light vehicle fuel economy to 24.1 mpg. The existing light
vehicle class gets only 22.1 mpg (2001) as gasoline mileage declines over time with
neglected maintenance.
Renewable energy use is slowly increasing (Chapter 13). Hydroelectricity usually
remains relatively constant, supplying 7.1% (2.8 quads) of US electricity, but it
dropped in the dry year of 2001 to 2.2 quads. Municipal waste and biomass fuels
generated 1.6% of electricity but it is growing at 1–2%/year. Wind energy con-
tributed 1.6 GWe of electrical production, but it is projected by EIA to grow 6%/year.
Geothermal contributes 0.3% of electricity but it is projected to grow 5%/year. So-
lar energy generated only 0.0003% of US electricity, but solar thermal energy is
projected to grow at 5%/year See Table 10.3.
Nuclear power’s role will diminish (Chapter 7). Nuclear power generated 20% of US
electricity in 2003, compared to France at 77% and Japan at 34%. Nuclear production
rose 15% over the past decade, not with new plants, but with a rise in capacity factor
from 71% to 88%. Nuclear energy is the second largest source of US electricity [after
coal (51%) and before natural gas (17%) and renewable (9.3%, hydro 7.1%) and
petroleum 3.0%]. EIA projects that electricity from natural gas is expected to grow
at 5%/year because of the attractiveness of combined cycle gas turbines. No new
nuclear reactors have been ordered in the United States in over 20 years, and there
4
Petroleum: transportation (13 Mbbl/day)/US-petroleum (20 Mbbl/day) = 65% in 2003.
Energy: transportation (26.9 quads)/US energy (98.2 quads) = 27%.
P1: GFZ
are no prospects for purchases of improved designs. EIA projects nuclear power
will grow to 8.4 quads/year in 2025 from its 2003 value of 8.0 quads/year. This
projection is difficult to comprehend since most reactors will be over 40-year old
and ready for decommissioning by 2025. In 2002 the Bush administration finalized
the decision to place spent nuclear fuel rods in the underground repository at Yucca
Mountain, but a July 2004 court decision disallowed the working limit of 10,000
years to consider environmental effects. Replacing the 100 GWe of US nuclear
power with combined-cycle gas-fired plants would require 4.5 TCF/year natural
gas, which is 20% of US consumption. Because of doubled natural gas costs and
uncertain future supplies from Canada, three utilities filed for site permits with the
Nuclear Regulatory Commission in 2003.
Stronger policies should be considered. The American Physical Society’s Panel on
Public Affairs (APS-POPA) recommended that the US government promote or
enact some combination of policies that would lower US consumption of oil. These
would include:
r higher gasoline taxes to reduce the use of oil

r higher fuel economy standards for passenger cars
r extension of car fuel standards to light trucks, minivans, SUVs
r development of a revenue-neutral “feebate” system of taxes and rebates to en-
courage purchase of vehicles with good fuel efficiency on a revenue neutral basis
r increased energy standards on buildings and appliances
r creation of a level economic playing field for all energy technologies
r adequate funding for research and development.
10.2 Energy Orders-of-Magnitude

10.2.1 Energy Per Person/Day
Daily per capita US consumption of fossil fuel is calculated from the following
factors:
r 300 million Americans consumed 98.2 quads (equivalent of 46 Mbbl/day)

r 86% of US energy obtained from fossil fuels
r use oil equivalent with 42 gal/bbl at 3.5 kg/gal.
Thus, per capital energy consumption in 2001 was
(47 Mbbl/day)(0.86 fossil)(42 gal/bbl)(3.5 kg/gal)/(300 M persons) = 20 kg/day,

(10.1)
or 44 lb/day. This corresponds to 30% the weight of a 140-pound person each day,
or 7.3 tons/year, or 600 tons in an 80-year life. The lifetime volume of a person’s
equivalent petroleum consumption is that of a 2500ft2 house.
P1: GFZ
10.2. Energy Orders-of-Magnitude 261
10.2.2 Kilowatthour
A 100-W bulb left on overnight for 10 hours consumes 1 kWh, corresponding to a
work equivalent of
W = 1 kWh = (1000 J/s)(3600 s) = 3.6 × 106 J. (10.2)
The work done by 10 c worth of electricity can raise an 80-kg body halfway up Mt.
Everest at 90% efficiency η, as calculated here:
h = ηE/mg = 0.9(3.6 × 106 J)/(80 × 9.8 J/m) = 4 km. (10.3)
The modern age depends on electricity made with modest fuel (1 kWh from
0.9 pound of coal) at cheap prices.5
10.2.3 Average and Peak Power

The United States consumed 3966 billion kWh in 2004, for an average power
Pavg = (3.996 × 1012 kWh/year)(1 year/8766 h) = 450 GWe , (10.4)
and average per capita power of 1.5 kWe (440 GWe /292 million persons).6 US
total capacity in 2004 was much larger at 935 GWe , which allows for a daytime
peak power 50% greater than average power, supplying a 10–15% reserve margin,
accommodating downtime for refueling and maintenance.
10.2.4 US Energy Use

Coal and nuclear power plants operate at 30–40% efficiency, but new combined-
cycle gas turbines (CCGT) using natural gas operate at 60%. At 33% efficiency, it
takes 4.5 thermal kilowatts (kWt ) to produce the per capita average electrical power
of 1.5 kWe . It took 39.6 quads to generate US electricity in 2003, which was 40% of
US total energy use. Combining these factors gives an average per capita thermal
power
Pavg/capita = (4.5 kWt )/(40% electrical) = 11.3 kWt . (10.5)
This is the heat power of 2–3 clothes dryers, operating full time over the entire year.
Total US thermal power is
PUS-thermal = (11.3 kWt )(292 M persons)(1 GWt /106 kWt ) = 3300 GWt . (10.6)
5
Prices in 2004: residences (8.9 c/kWh), commerce (8.0 c/kWh) and industry (5.3 c/kWh).
On average, generation costs 5.0 c/kWh, while transmission costs 0.5 c/kWh and distri-
bution costs 2.1 c/kWh. California (2006) residential costs rise with use in five brackets of
11.4 c, 13.0 c, 23 c, 32 c and 37 c.
6
Average US electrical power was as follows in 2004: coal (219 GWe ), nuclear (90.0 GWe ), nat-
ural gas (55.3 GWe ), hydroelectric (30.1 GWe ), petroleum (12.5 GWe ), geothermal (1.6 GWe )
and wind (1.6 GWe ).
P1: GFZ
10.2.5 About Units

Practical energy units are an impediment to a good understanding of energy issues
(see Appendix D), but since they are used in real-life discourse, we come to terms
with them here. We begin by putting in context the unit “million barrels of oil
per day” (Mbbl/day, but some reference it as mbod). US power consumption is
expressed in a variety of units:
P = (3.3 × 1012 W)(8.64 × 104 s/day)(1 J/W-s) = 2.9 × 1017 J/day (10.7)
P = (2.9 × 1017 J/day)(1 Btu/1055 J) = 2.7 × 1014 Btu/day (10.8)
P = (2.7 × 10 Btu/day)(1 bbl/5.8 × 10 Btu) = 47 Mbbl/day
14 6
(10.9)
P = (2.7 × 10 Btu/day)(365 day/year)(1 quad/10 Btu) = 99 quads/year.
14 15
(10.10)
These estimates agree with US consumption of 98.2 quads in 2003. Other energy
units are the exajoule (EJ = 1018 J), the terawatt-year (TWyr = 8.77 × 1012 kWh),
1 ton of oil (39.7 MBtu), trillion cubic feet of natural gas (1 TCF = 1 quad), and
Mbbl/day (2.12 quad/year).
10.3 Fossil Fuel Models

10.3.1 Exponential Growth
Domestic oil production increased at λ = 9% per year after its 1860 discovery in
Titusville, Pennsylvania until 1920. The rate of increase then declined from 1920 un-
til the peak production year 1970, when oil production in the lower 48 states began
to diminish. Before the oil embargo US total energy consumption grew at 4.3%/year,
but after the oil embargo energy growth dropped to 1.5%/year. Reduction in growth
was a result of enhanced end-use efficiency and a change in societal working modes.
Enhanced end-use efficiency comes from better engineering, while conservation is
exemplified by the wearing of sweaters to keep warm and by choosing walking
over driving.
The rate of energy consumption, a power such as quads per year, usually in-
creases exponentially:
d E/dt = P = Po e λt , (10.11)
where Po is initial power at t = 0 and λ is growth rate. The total energy consumed
E in a time interval 0 to tf is the time integral of power,
tf
E = E f − E o = Po e λt dt = (Po /λ)[e λt f − 1]. (10.12)
0
If the time interval is long, that is λtf 1, cumulative energy increases expo-
nentially, as Po e λt f /λ. For short times, λtf 1, the growth rate λ is the fractional
P1: GFZ
10.3. Fossil Fuel Models 263
change in power (or cumulative energy) divided by the time interval,

λ = (P − Po )/Po t f = (P/Po )/t f (10.13)
Longer time intervals require the use of logarithms to obtain λ:
λ = ln[Pf /Po ]/t. (10.14)
It is foolhardy to rely on exponential functions to project future energy use. No
scientific law dictates their correctness. Sometimes these functions work, but often
they are not accurate, and sometimes they are grossly inaccurate, as it the case for
energy planning in 1970. The 4.3%/year growth rate for total energy, applied to
1973 embargo year of 76.6 quads, gives a bad prediction indeed for consumption
in 2003:
P2003 = P1973 e λt = (76.6 quads/year)e (0.043 × 30) = 76.6 × 3.19 = 278 quads/year.
(10.15)
This overestimates reality by a factor of 2.8 since actual energy use in 2004 was 99.7
quads. A 1% growth rate gives a more reasonable estimate:
P2000 = (76.6 quads/year)e (0.01 × 30) = 76.6 × 1.31 = 103 quads/yr. (10.16)
The 1% rate was in error in that it did use the temporary reduction to 70 quads/year
in the early 1980s.
A quick approach to making predictions with exponential growth is to determine
TD , the time needed, to double Po :
2Po = Po e λTD . (10.17)
Rearranging gives 2 = exp(λTD ). Taking the ln of both sides gives
ln[2Po /Po ] = ln 2 = 0.693 = ln[exp(λTD )], or (10.18)
70 = λTD (10.19)
with λ in percent/year and TD in years. US electricity, which grew at 7%/year in
the 1960s and early 1970s, had a doubling time of TD = 70/7 = 10 years. If we use
7%/year growth for US electrical power, the average power of 200 GWe in 1970
would have grown to 400 GWe in 1980, 800 GWe in 1990 and 1600 GWe in 2000. As
it happened, the 1600 GWe projection for 2000 is four times the 2001 average of 403
GWe .
10.3.2 Pedagogical Petroleum Model

Neils Bohr commented that there is nothing more difficult to predict than the future.
However, King Hubbert successfully predicted in 1957 that US oil production from
the lower 48 states would peak between 1966 and 1971, and he correctly bracketed
peak production of 9.1 Mbbl/day. in 1970. Hubbert’s model was successful in
times of constant oil prices, but it was not successful beyond 1980, as it does not
take into account the dismal science of economics or new modes of oil production
P1: GFZ
1 Figure 10.14. Lower 48 states

petroleum production, 1860–1920.
The log of production rate in
0 Mbbl/day is plotted versus the
year of production. Data were
fit to reveal a growth rate of
-1 λ = 0.09/year. The first year’s
production (1859) at Titusville,
Pennsylvania, was not included in
-2 the data fit.
-3
1840 1860 1880 1900 1920 1940
from offshore. Chapter 16 extends Hubbert’s model to include economics, a good

next step. However, theories also have to include the realities of societal changes,
alternate fuels, improvements in efficiency and enhanced petroleum recovery.
Figure 10.14 shows that US petroleum production had exponential growth dur-
ing its first six decades (1860–1920). The excellent linear fit between the log of
petroleum production and time gives a phenomenological link between production
rate Q̇ and cumulative petroleum production Q:
Q̇ = λQ = λQo e λt , (10.20)
where Qo is initial cumulative production at t = 0. The solution Q = Qo e λt sug-

gests production rate Q̇ is proportional to the maturity of the petroleum industry,
represented by total petroleum produced Q. Fitting the 1860–1920 data, we obtain
λ = 0.090/year (9% per year).
The 9% growth rate does not apply to production after 1920 since at that time
supply showed some limits. Hubbert modified the exponential rate equation in
two ways. Hubbert defined the ultimate petroleum available at a given production cost
as Q∞ . He further required the production rate dQ/dt (or Q dot) be proportional to
the fraction of unused resource (1 − Q/Q∞ ). Production stops (dQ/dt = 0) when the
resource is depleted (Qfinal = Q∞ ) as is seen in Hubbert’s equation:
d Q/dt = λ(Q − Q2 /Q∞ ) = λ(Q/Q∞ )(Q∞ − Q) = k Q(Q∞ − Q) (10.21)
where λ is the initial growth rate and k = λ/Q∞ . This equation is the Verhulst
equation used in condensed matter physics, which has solutions similar to Gaus-
sian curves. The production rate symmetrically rises and falls, centered at the year
of maximum pumping. In the early oil days, when cumulative production Q is
small, the model simplifies to pure exponential growth, dQ/dt = λQ. The solu-
tion Q = Qo e λt and its logarithm, ln(Q/Qo ) = λt, are consistent with the data in
Fig. 10.14.
P1: GFZ
Figure 10.15. US petroleum pro- 10

duction from lower 48 states, 1860–
1995. Petroleum production is in
8
Mbbl/day. The area under the full
curve gives Q∞ = 150 Gbbl and λ
= 0.090 per year. 6
0
1860 1880 1900 1920 1940 1960 1980 2000
Lower-48 production from 1860 to 1980 (120 years) is reasonably symmetrical

around the 1970 maximum. A numerically integrated solution shown in Fig. 10.15
fits the data reasonably up to 1980, but it then fails as production remains at about
one-half the maximum level (as offshore production increases). To determine the
ultimate resource Q∞ we take the derivative of the production rate with respect
to Q and set it equal to zero. The top of the resource curve with a zero derivative
takes place at the maximum pumping rate:
[d(d Q/dt)/d Q] Qmax = 0 = k(Q∞ − 2Qmax ). (10.22)
At this point one-half the resource has been consumed, Qmax = Q∞ /2. Thus,
d Q/dtmax = k Qmax (Q∞ − Qmax ) = kQ∞ 2 /4. (10.23)
Solving for Q∞ using the 1970 value of dQ/dtmax = 9.1 Mbbl/day = 3.3 Gbbl/year
gives
Q∞ = 4d Q/dtmax /λ = (4)(3.3 Gbbl/year)/(0.09/year) = 147 Gbbl. (10.24)
This is fairly close to Hubbert’s 1969 estimate of 165 Gbbl, but smaller than the 195
Gbbl already produced by 2004. Using Q∞ = 147 Gbbl and λ = kQ∞ = 0.090/year,
we obtain the curve in Fig. 10.15. The area under the curve is the complete resource
Q∞ , which is within 6% of (dQ/dtmax )(2), which is the product of the maximum
pumping rate, dQ/dtmax , and twice the half-width in time at half-maximum, 2.
Our estimate of the mature lifetime of US petroleum is
2 = Q∞ /(d Q/dtmax ) = Q∞ /(k Q∞ 2 /4) = 4/k Q∞ = 4/λ = 4/0.09 = 44 years,

(10.25)
which is close to Hubbert’s value of 45 years. Our estimate is too small by perhaps
a factor of 2, because higher prices, offshore oil and new technologies increase the
size of the resource.
P1: GFZ
These results can be compared to estimates by the US Geological Survey. In

2000 USGS estimated the United States had 30.5 Gbbl of proved reserves (EIA gave a
figure of 21.4 Gbbl in 2004), 76.0 Gbbl of reserve growth and 83.0 Gbbl of undiscovered
resources. The 2006 USGS global estimate was 1293 Gbbl (63% in the Middle East)
of proved reserves, 730 Gbbl of reserve growth and 939 Gbbl of undiscovered resource.
Higher prices give extra money to drill higher-cost resources and use higher-
cost recovery methods, increasing the recoverable resource Q∞ . In addition, higher
petroleum prices reduce petroleum demand. It is generally agreed that relatively
small increases in price have a small effect on demand since initial demand elasticity
and price rise are both small. However, a price rise of several dollars per gallon
reduces demand considerably beyond that predicted with the initial choice for
elasticity. This is because the magnitude of demand elasticity increases for larger
price rises. Section 16.6 modifies Hubbert’s model to include a supply and demand
elasticity that reduces demand and expands supply resource Q∞ .
10.3.3 Natural Gas

In times of energy shortages societies have used gas recovery systems. Europe made
producer gas for automobiles in World War II by burning charcoal in steam and
limited air at 1000◦ C. Producer gas contains partially consumed carbon monoxide
(20%), as well as hydrogen (14%) and methane (4%), and noncombustible gases.
Producer gas contained only 14% the energy of natural gas’s 1000 Btu/ft3 (mostly
methane), but the cars ran. The world is betting on better times to bring copious
supplies of natural gas, as the bridging fuel to the future.
Natural gas consumption. Hubbert’s model has been less successful with natural
gas and coal production. Natural gas consumption dropped 30% after the oil em-
bargo, primarily because electricity production from natural gas was curtailed in
favor of coal.7 However, natural gas was expected to increase more than any other
fuel, rising by 50% from 22.5 TCF in 2003 to 35.8 TCF in 2025, but this is no longer
believed. The wide adoption of CCGT was expected to considerably increase nat-
ural gas consumption. The United States imported 3.5 TCF of natural gas, mostly
from Canada, in 2003 (15% of US consumption), with EIA projecting that it would
grow to 5 TCF/year by 2025. However, it was announced in June 2003 that Canada
will not be able to markedly increase its imports to the United States until 2008.
This report and other factors drove US prices from $2.60 per thousand cubic feet
for utilities in 1999 to $6 per thousand cubic feet in 2005. These compelled utilities
to reconsider coal for 100 new power plants. Liquefied natural gas (LNG) imports
are projected (2006) to rise from 0.5 TCF in 2003 to 4.3 TCF in 2025. This requires
substantially increasing US LNG port facilities. The safety factor is prompting im-
porters to develop offshore landing facilities. The best bridging fuel is now more
uncertain in supplies and costs.
7
EIA natural gas percentiles: electricity (25%), industry (32%), residences (22%), commercial
(14%), other (7%).
P1: GFZ
Natural gas reserves. The US proven reserve is 193 TCF (EIA, 2006), but
itstechnically recoverable natural gas resource is much larger, optimistically at 1300
to 1400 TCF8 , an amount that could supply present US use of 23 TCF/year for
50 years without growth, or 35 years at 1.8%/year growth. It is not clear how much
of the technically recoverable resources can be converted to proved reserves. Some
reserves are at depths of over 3000 m and some are located in difficult places.
Besides these hindrances natural gas often contains unwanted gases that must
be removed. Because the gas industry is fragmented, it is difficult to coordinate
a shift from technically recoverable resources to proved reserves. US proved reserves
dropped from 200 TCF in 1983 to 193 TCF in 2006. World proved reserves are 6112
TCF (2006). Predominant proved reserves are in Russia (1680 TCF), Iran (971 TCF),
Qatar (911 TCF), Saudi Arabia (241), United Arab Emirates (214 TCF), and the
United States (193 TCF). These numbers total 69% of the world’s 5500 TCF proved
reserves.
10.3.4 Coal
The United States consumed 1125 Mtons of coal and exported 21 Mtons in 2004,
with 92% for electrical generation and 8% for industry. Coal is a less versatile
energy source than petroleum and natural gas since mining and burning are en-
vironmentally harmful. Mining is moving from the East (45%) to the West (55%)
and from underground mining (30%) to surface mining (70%). In 2003 the average
price of coal was $17/ton with anthracite at $50 and bituminous at $30. As natural
gas prices doubled during 2000–05, coal revived from essentially no plans for new
coal-powered plants to plans for 100 additional plants.
The United States has a recoverable resource of 270 billion tons (270 Gton in
2003, but perhaps as much as 1000 Gton), which can last for 270 years with
no growth at 1 Gton/year. However, the impact of mining and burning con-
strains coal’s projected growth to about 1%/year, lowering the lifetime to 50 years.
After the oil embargo, many electrical power plants converted from petroleum
and natural gas to coal, with the result that coal produced 2.5 times as much
electricity as oil and gas together in 2003. The expansion of coal (from 13 to
22.7 quads/year) fulfilled 50% of US energy growth since the oil embargo. Sul-
fur dioxide emissions decreased because of the adoption of scrubbers and the
shift to low sulfur coal from west of the Mississippi. Nevertheless, acid deposi-
tion from mine runoff and air pollution continues to cause environmental prob-
lems and air pollution deaths (Chapter 6). Coal also is the most intensive pro-
ducer of carbon dioxide because it is almost pure carbon (not much hydrogen)
and because coal is burned at lower power plant efficiencies than natural gas in
CCGTs.
8
EIA natural gas total resource estimate of 1300 TCF (2003): proven (184 TCF), undiscov-
ered nonassociated conventional (269 TCF), inferred nonassociated conventional (222 TCF),
unconventional (445 TCF), other (169 TCF).
P1: GFZ
10.4 Energy Rates of Change

The ratio of national energy use E to gross domestic product GDP is not constant
since it decreases with improved end-use efficiency, and it ignores shifts between
different fuels. In addition, the E/GDP ratio does not take into account modal shifts
in the economy, which changed from 1980 employment [services (66%), manufac-
turing (22%), agriculture (4%)] to a more service-oriented economy [77%/15%/2%].
As steel mills were replaced with computer software companies, energy use de-
clined while GDP rose. Before the oil embargo, the E/GDP ratio was fairly constant,
implying lockstep coupling between the economy and energy use. After the oil em-
bargo, the ratio decreased from 18,000 Btu/$ in 1973 to 11,000 Btu/$ in 1999 (1999
dollars). Perhaps 50% of the reduction resulted from increased end-use efficient
automobiles, appliances and buildings, with the remainder coming from modal
shifts in the economy, that is, the movement of workers between labor sectors.
The basic variables needed to analyze national energy use—elements that give a
more complete picture—are gross domestic product (GDP), “efficiency” of turning
energy into money (η = GDP/E in $/Btu), and per capita productivity of the
labor force (Prod = GDP/L). Efficiency η takes into account both enhanced end-use
efficiency and modal shifts within the economy. Using these variables, the national
energy use on an annual basis is
E = GDP/η = ProdL/η. (10.26)
For small changes, its differential is
E/E = GDP/GDP − η/η = Prod/Prod + L/L − η/η. (10.27)
Over small periods of time, growth in productivity and growth in the number
of laborers are additive, while growth in efficiency is subtractive. For example, in
the years since the oil embargo, the GDP grew 3.0% annually (in constant dollars)
while US population grew at 1.1% and productivity grew at 1.9%. Since energy
grew at 1.1%, efficiency increased at 1.9%. For the 1973–86 time frame, energy con-
sumption remained essentially constant at the 1973 and 1986 endpoints, the 2.8%
GDP growth was due to improved efficiency and modal shifts. This improvement
is not surprising, since formerly the United States had abundant energy supplies
and was less concerned about energy efficiency.
In 2003 EIA projected 1.5% growth in energy and continued increase in efficiency
of 1%. This gives a rise in GDP of 2.5%, from 1.5% growth in productivity and
1% growth in population. National adoption of energy efficiency in buildings,
transportation, and industry can lower the 1.5% energy rate. See Table 10.4.
Table 10.4. Annual rates of change in productivity,

laborers, efficiency, and energy (EIA data)
Prod/Prod + L/L −η/η = E/E
1973–2000 1.9% 1.1% −1.9% = 1.1%
1973–1986 1.8% 1.0% −2.6% = 0.2%
1990–2000 2.0% 1.2% −1.6% = 1.6%
P1: GFZ
10.5. Population and Sustainability 269
10.5 Population and Sustainability

Population drives future energy needs. From year 0 to 1650, Earth’s population
doubled, going from 250 million to 500 million at a slow rate of 0.04%/year. Global
population growth peaked in 1965–70 at 2.1%, dropping to 1.6% in 1995 and 1.2%
in 2002 with a population of 6.3 billion. Over several years the United Nations
has revised its population estimates for 2050 from over 10 billion to 8.9 billion,
taking into account greater use of birth control, the spread of HIV/AIDS, improved
economies and improved status of women. The global average number of births per
female dropped from 5 in 1950–55 to 2.7 in 2000–05, and it is projected to drop to 2.1
in 2050. The industrialized nations had an average birth rate of 1.58 births per female
in 2002, far below the 2.1 replacement value. In 1960, five countries had fertility
rates below 2.1, while in 2000 there were 64 countries below 2.1, representing 44%
of the global population. The average in lesser developed countries was 3.11 births
per female. The values for the continents (2002) were Africa (5.22), Asia (2.72),
Europe (1.42), Latin America (2.72), and North America (2.01).9
In 1980 China began a plan that would lower its population from 975 million
to 700 million. The decrease was to be achieved through coercion, for example,
through forced abortion. The reduction did not take place, but China’s population
growth did lessen. The UN projects China’s population will stabilize at 1.4 billion
by 2050. India has a more relaxed view on population and will pass China in
about 2030. US population growth was 1.09%/year during 1995–2000, growing at
3 million/year, one-half of this from immigration. About 10% of United States is
foreign born, about the same as its historic average between 1850 and 1995. Japan,
the world’s second largest economy, had a birthrate of 1.39 births per female in 2002,
which projects to a reduction in population from 128 million in 2002 to 110 million
in 2050. Germany, the third largest economy, had a birthrate of 1.34; its population
is projected to go from 82 million in 2002 to 79 million in 2050, moderated by
immigration. UN estimates (Table 10.5 and 10.6) include immigration and the use of
cohort population groups (discussed below). The UN projects the global population
of persons over 65 years of age will rise from 19% of the total population in 2002
to 32% in 2050.
The rate of change in population can be written as
d P/dt = b P − d P (10.28)
where b and d are the birth and death rates on a per capita basis. This differential
equation has a simple solution for short time frames, P = Po e (b−d)t , but the parame-
ters b and d are not constant. A more accurate approach is to describe the population
in terms of cohorts of similar age. Cohorts change with time as birth/death rates
vary for each cohort (problem 10.18). This approach is fairly accurate for 20-year
projections since living persons dictate the results.
9
US replacement values have varied over time: 1920 (3.333), 1955 (3.767), 1975 (1.738), 2000
(2.130). The 2000 data by population group was Hispanic (3.11), black (2.19), white (2.11).
[National Center for Health Statistics, 2003]
P1: GFZ
Table 10.5. UN population data and projections (United Nations Population

Department, February 2003)
population (millions) %/yr %/yr births/female
1950 2002 2050 1995–2000 2045–2050 2002

World 2519 6301 8919 1.35 0.33 2.83
China 555 1304 1395 0.90 −0.37 1.80
Germany 68 82 79 0.15 −0.15 1.34
India 358 1065 1531 1.76 0.26 3.45
Japan 83 128 110 0.25 −0.56 1.39
Mexico 28 99 140 1.64 0.03 2.75
Pakistan 40 154 349 2.62 0.99 5.48
Russia 103 143 101 −0.34 −0.86 1.25
UK 50 59 66 0.34 0.12 1.70
US 158 284 408 1.09 0.41 2.05
Thomas Malthus concluded in 1798 that “[t]he happiness of a country . . . . [de-

pends] upon the degree in which the yearly increase of food approaches to the
yearly increase of an unrestricted population . . . . The power of population is in-
definitely greater than the power in the Earth to produce subsistence for man.”
Malthus failed to predict technological progress, as industrialized nations have
gone on to feed their populations with a small work force. However Malthus may
ultimately be proven correct, as Earth has other limitations. Another approach takes
into account the carrying capacity of Earth. Analyst Joel Cohen modeled population
growth dP/dt by adapting the Verhulst equation (Section 10.3) to
d P/dt = r P[K − P] (10.29)
where r is the Malthusian growth constant and K is the carrying capacity of Earth.
Cohen increases carrying capacity as progress accompanies population growth. Yet,
higher population densities can hurt sustainability of animal species that depend
on sufficiently large ecosystem areas. Predator species are extinguished first when
ecosystems become smaller. Without natural predators, prey species became more
abundant. Another consideration is that half of the world’s employment depends
Table 10.6. US GDP, population, laborers, and energy use (1960–2000)

Year GDP Population Workers Energy Global population
1950 $1687 151 — 34.6 2555
1960 $2377 179 70 45.1 3030
1970 $3578 203 83 67.9 3708
1980 $4901 227 107 78.4 4457
1990 $6708 249 126 84.6 5284
2000 $9224 281 141 99.3 6080
GDP in billions of 1996 dollars, population and workers in millions, energy in

quads/yr (EIA, 2003)
P1: GFZ
10.6. Single and Combined Cycle Power Plants 271
on small-scale farms, forests and fisheries. However, two-thirds of the fisheries are
threatened by overharvesting and pollution, and one-third of Earth’s people are
confronted with water scarcity. Energy use is tied to these issues in complicated
ways, both positive and negative.
10.6 Single and Combined Cycle Power Plants

Electricity generation with combined cycle gas turbines (CCGT) is the best supply
side advance since the oil embargo. A typical steam power plant requires about
10,000 Btu (10.55 MJ) to make 1 kWhe of electricity, which contains energy in the
amount
1 kWhe = (1000 W)(3600 s) = (3.6 MJ)(1 Btu/1055 J) = 3412 Btu. (10.30)
Hence, the efficiency of a typical power plant is 30–40% (3.6 MJ/10.5 MJ). How-
ever, General Electric’s 0.5 GWe CCGT needs only 5690 Btu (6.0 MJ) of natural gas to
produce a kWh with 60% efficiency (3.6 MJ/6.0 MJ) for a 43% savings. If the United
States switched one-half its grid (215 GWe of coal and 14 GWe of petroleum) to
CCGTs, there would be a savings of
43%(0.5 × 37.7 quads/year electricity) = 8 quads/year, (10.31)
which is 8% of US energy use. CCGTs produced 3% of US electricity in 1999, but
they comprise 88% of planned power plants. This makes sense since CCGTs can be
erected in a year at a cheap price of 40 c/We to make electricity at 2–3 c/kWh at the
busbar. CCGTs use less space and pollute much less than coal plants. CCGTs emit
60% less CO2 (0.10 kg C/kWh) than a 40% efficient coal plant (0.26 kg C/kWh).
The United States was betting that natural gas would be plentiful for a transition to
CCGTs. However, Canada now appear less certain for increased gas supplies and
the cost of gas has doubled during 2000–05, halting the shift to CCGTs.
10.6.1 Rankine Cycle

Steam power plants operate on the Rankine cycle, which is a continuous water
cycle of pressurization, vaporization, expansion, and condensation in a sealed loop.
Carnot cycles cannot be practically applied to power plants, though they provide
a theoretical start. The efficiency of a Carnot heat engine is
ηCarnot = Wout /Qhot = 1 − Qcold /Qhot = 1 − Tcold /Thot = T/Thot , (10.32)
where T = Thot − Tcold is the temperature difference between the hot and cold
reservoirs. The Rankine cycle is used in power plants because (1) it does not have
the Carnot isothermal expansions and compressions which are exceedingly slow
and (2) because water has vaporization and condensation phase transitions. The
Rankine cycle efficiency is
ηRankine = (h 6 − h 5 )/(h 5 − h 1 ), (10.33)
P1: GFZ
Figure 10.16. Rankine cycle with superheat. The Rankine cycle pressure vs. volume diagram
begins with a pump raising the pressure of water at constant volume to enter the boiler (1–2).
The boiler heats at constant pressure to convert water to steam, raising its temperature and
expanding volume (2–4). A superheating boiler further raises the temperature of steam (4–5).
The high–pressure steam adiabatically expands, spinning a turbine (5–6). Steam is condensed
in a heat exchanger, after spinning the turbine, at constant low pressure to complete the cycle
(6–1).
where enthalpy h is the available energy, that is, h = internal energy + pressure times
volume = U + pV. The subscripts refer to Fig. 10.16, where h i is enthalpy at stage i of
the cycle. Enthalpy values are readily obtained from steam tables and steam-dome
charts.
A Rankine cycle with a condenser temperature of 39◦ C (102◦ F) and a boiler
temperature of 336◦ C (636◦ F) at 2000 psi has a theoretical efficiency ηRankine = 39%,
compared to ηCarnot = 49%. By increasing steam to 538◦ C (1000◦ F) with a superheat
cycle, ηsuperheat rises to 43%. An additional 1% rise in efficiency can be obtained
with a reheat cycle that superheats steam part way through the expansion (4–5) to
obtain ηsuperheat-reheat = 44%. Operating Rankine systems operate at η = 30–40%.
10.6.2 1-GWe Coal Plant

The energy content of coal varies from 8,000 to 12,000 Btu/pound with a US average
of 10,400 Btu/lb. It takes a little less than 1 pound of coal to make 1 kWh, since
it takes on average 10,000 Btu to make 1 kWh.10 To sustain a typical US life at
an average power of 1.5 kWe requires 36 pounds of coal/day (1.5 kW × 24 h =
36 kWh), which is 6.6 tons/year or 500 tons for an 80-year life.
A 1-GWe plant operating 24 h produces about 24 million kWh/day from 20 mil-
lion pounds of coal, or 10,000 tons/day. It takes a daily train of some 100 freight
10
At 40% efficiency, it takes 3412 Btu/0.4 = 8530 Btu of coal, which is 0.7 lb of 12,000 Btu/lb
coal, or 1.1 lb of 8000 Btu/lb coal. If efficiency is 30% it takes 0.95–1.4 lb.
P1: GFZ
10.6. Single and Combined Cycle Power Plants 273
cars, each carrying 100 tons of coal, to operate the plant. At efficiency η = 33%, the
rejected thermal power is 2 GWt , (1.7 GWt in coolant and 0.3 GWt in stack gases).
To heat water 10◦ C (18◦ F) requires a water flow of
dm/dt = 1.7 GWt /cT = (1.7 × 109 W)/(4200 J/kg-K)(10 K) = 4 × 104 kg/s
(10.34)
10.6.3 The Brayton Cycle

Gas turbines are a mature technology, used in jet aircraft and M-1 tanks. Rankine
steam cycles fail to take advantage of the high temperatures reached by combustion
gases because of high steam pressure. However, combustion gas densities in gas
turbines can be smaller, allowing higher temperatures at smaller pressures than
steam cycles. Brayton cycles typically operate at 1200◦ C (2200◦ F), considerably
above the steam-based Rankine cycles (540◦ C, 1000◦ F), but at 1.5 MPa (225 psi)
pressure, far below Rankine’s 14 MPa (2000 psi). A Brayton open-cycle turbine
uses fresh air, exhausting to the atmosphere with a theoretical efficiency over 50%,
but in practice η = 30–40%. See Fig. 10.17.
10.6.4 Combined Cycle Gas Turbines

CCGTs use the hot exhaust from the turbine to produce steam in a heat exchanger
for a second-stage cycle. Modern CCGTs obtain ηCCGT = 60% by obtaining an extra
boost by vaporizing additional steam in the hollow turbine blades in the first stage
for use in the second stage. The efficiency of a CCGT is
ηCCGT = (WB + WR )/Qin = ηB + WR /Qin , (10.35)
where WB and WR are the work produced in the Brayton and Rankine cycles, re-
spectively, and Qin is heat energy entering the Brayton cycle. The exhaust heat of
p
QH
p2 2 3
Figure 10.17. Open Brayton

cycle. The Brayton cycle con- Tc
sists of adiabatic compression
of gas (1–2); heat exchanger
and gas combustion at constant
pressure (2–3); adiabatic expan- Ta
sion in a turbine (3–4); and
a heat exchanger at constant p1 4
pressure to preheat the enter- 1
ing fuel/air mixture, or to make V
Qc
steam for a second cycle (4–1).
P1: GFZ
the Brayton cycle is the input heat of the Rankine cycle, giving WR = ηR (1 − ηB )Qin ,
for a CCGT efficiency of
ηCCGT = ηB + ηR (1 − ηB ). (10.36)
Using a first-stage efficiency of 40% and a second stage efficiency of 30% gives
ηCCGT = 58%, which is close to the 60% obtained by modern CCGTs.
10.7 LNG Explosions

The aftermath of the September 11, 2001, terrorist attack temporarily closed US
liquefied natural gas (LNG) ports because of the possibility of huge explosions at
the four US LNG ports. The January 2005 billion-dollar LNG explosion at the port of
Skikdo, Algeria, heightens these concerns. This is a large issue since 50 LNG ports
are being planned around the world to enhance natural gas supplies. Unloading
LNG ten miles from land is being considered to lessen these concerns. In the past,
Middle East oil fields were flared, disposing of natural gas by wastefully burning it.
This is no longer true, as LNG has become economically viable. LNG ships carry up
to five 35-m diameter spherical tanks, each holding 22,400 m3 of LNG. If the tanks
rupture, LNG vapor would spread horizontally, rather than rise, because cold LNG
vapor (−160◦ C) is heavier than air. Our estimate for the explosive energy of LNG
is a worst case scenario since spreading, air-mixing and burning are assumed to be
complete. The energy content of one LNG tank is
E = (2.24 × 104 m3 )(3.3 × 1010 J/m3 ) = 8.2 × 1014 J. (10.37)
This is equivalent to an explosive yield
E = (8.2 × 1014 J)(1 kton/4.2 × 1012 J) = 200 ktons, (10.38)
which is 15 times the Hiroshima bomb. If all five tanks exploded, the yield would
be 1 Mton. The duration of the explosion would depend on weather and mixing.
If LNG migrated horizontally a considerable distance, it might take 5 minutes to
burn, giving a thermal power
Pthermal = E/t = (8.2 × 1014 J)/(300 s)(1 W/J-s) = 2700 GWt (10.39)
for one tank and 13,500 GWt for five tanks. Considerable discussion took place on
the likelihood of explosions when LNG ports were first proposed, but, thus far, the
safety record has been excellent. The risk probability for accidents that kill more
than one person was estimated in the 1980s at 2 × 10−5 to 5 × 10−7 per tanker
transit.
Problems
10.1 Daily shower. A shower uses 10 gallons of water heated from 60◦ F to 100◦ F
with electricity from coal of 104 Btu/lb at η = 40%. How much coal is needed
for one shower and for 80 years of daily showers? How many acre-feet/year
P1: GFZ
Problems 275
of water are needed to clean coal for a lifetime’s showers if the weight ratio
of coal/water = 0.5? What does a shower cost at 8.6 c/kWh? (A square mile
has 640 acres and a gallon of water weighs 8.3 pounds.)
10.2 Clothes drier. A 5 kW clothes drier is used for two hours. How much coal
at 35% efficiency was burned in the process? How much natural gas at 60%
efficiency was burned?
10.3 Global energy. The world consumed 346 quads in 1990 and 412 quads in
2002. What are the linear and fractional growth rates? What will the world
consume in 2100 under linear and exponential growth?
10.4 Energy growth. The US consumed 98.2 quads in 2003. If energy use increases
(a) 2 quad/year, (b) 2%/year, what will energy use be in 2100 and how much
will be consumed during the century?
10.5 Binding energy. Heptane (C7 H16 ) contains 1.15 kilocalories/100 grams. What
is heptane’s eV/atom binding energy?
10.6 US oil/gas. (a) US natural gas consumption is projected to rise 1.8%/year
from a 2003 level of 22.5 TCF. If only US resources are used, how long will
it take to exhaust proved reserves (2002) of 184 TCF and technically recoverable
resources of 1300–1400 TCF? (b) US petroleum consumption is projected to
rise 1.6%/year from a 2003 level of 20 Mbbl/day. If only US resources are
used, how long will it take to exhaust US proved reserves of 22.7 Gbbl and
technically recoverable resources of 144 Gbbl?
10.7 Bridging fuel. (a) What is 50% of Carnot efficiency of heat engines at inputs
of 600◦ C (1200◦ C) with 40◦ C exhaust? (b) How much natural gas is annu-
ally consumed by a 1-GWe plant at η = 40%, with a load factor of 0.85 and
1000 Btu/ft3 ? (c) What is annual consumption if η = 60%? (d) How much gas
would be needed to replace global nuclear power plant capacity of 350 GWe
(load factor of 0.85) with CCGTs? How does this compare to global natural
gas consumption of 95.2 TCF/year in 2002? (e) What is the growth rate for
the EIA global projection of 176 TCF in 2025?
10.8 Global oil/energy. (a) The Oil and Gas Journal estimated global proved-
petroleum reserves at 1200 Gbbl in 2003. How long will oil last at the projected
increase of 2.3%/year (mostly in Asia) from the base of 80 Mbbl/day in 2003.
How long will oil last if the “technically recoverable resource” of 5000 Gbbl
is viable? (b) The world consumed 412 quads of fossil fuels in 2002 with
a growth of 2%/year. How long will fossil fuels last from a global 280,000
quads of conventional and unconventional, reserves and resources, which
are dominated by coal?
10.9 First and second laws. (a) Is it correct to claim a heat engine obtains η = 32%
from a 200◦ C source? (b) What is the second law efficiency of a 90% efficient
gas heater?
10.10 Cooling water. A 1-GWe coal plant loses 65% of input heat, divided into 12%
stack emissions and 53% rejected into the ocean. (a) If discharge cooling-
water is 10◦ C warmer than the intake water, what is the water flow rate? (b)
What is the cooling-water flow rate from a 1-GWe nuclear plant with η =
32% and no stack emissions?
P1: GFZ
10.11 Strip mining. (a) A coal seam is 20 meters thick with a specific gravity of 1.3.
What area must be strip mined annually to supply a 1 GWe plant with η =
40% and an 80% load factor? (b) How much area must be annually stripped
if all US production of 1109 million tons in 2001 was obtained this way?
10.12 Hubbert petroleum model. (a) Show that the lifetime of petroleum resource
is approximately 2 = 4/λ (Eq. 10.25). (b) Obtain and plot the historic
production rate for natural gas from www.EIA.gov. Does Hubbert’s model
describe natural gas production?
10.13 $/M Btu. What are fuel prices in $/MBTU: petroleum ($30/bbl, $2 to $5 per
gal), natural gas ($6 per 1000 ft3 delivered), coal ($40/ton delivered) and
electricity (8.6 c/kWh, 2001 residential)?
10.14 US population. US population was 106 million in 1920, 123 M (1930), 132 M
(1940), 151 M (1950), 180 M (1960), 205 M (1970), 227 M (1980), 249 M in 1990,
276 M in 2000 and 292 M in 2003. Plot these data on linear and semilog graphs.
What is λ growth rate each decade? How has the doubling time changed over
the decades? What is your projection for US population in 2100? How does
the 2004 UN projection of 420 million in 2050 fit the curve?
10.15 Earth population. What will Earth’s population be in 2100 if it was 6.3 billion
in 2003 and growing at 1.2%/year? How long would it take for the total mass
of Earth’s people (60 kg each) to equal 50% of Earth mass (6 × 1024 kg)?
10.16 GDP/energy. (a) Determine US productivity (GDP/laborer) and efficiency
η(GDP/energy) as a function of time from the data below. (b) Determine US
decade growth rates for productivity, η, population (US/global), work force
and energy used.
US GDP, population, laborers and energy use (1960–2000) GDP in billions
of 1996 dollars, population and workers in millions, energy in quads/year
(EIA, 2003).
Year GDP Population Workers Energy Global population

1950 $1687 151 — 34.6 2555
1960 $2377 179 70 45.1 3030
1970 $3578 203 83 67.9 3708
1980 $4901 227 107 78.4 4457
1990 $6708 249 126 84.6 5284
2000 $9224 281 141 99.3 6080
10.17 China/India. China’s 2002 population of 1.304 billion grows at 0.9%/year,

while India’s population of 1.065 billion grows at 1.76%/year. How do these
rates of growth relate to the birth rates of 1.80 births/female for China and
3.45 for India? What will be the populations of these countries in 50 years?
When will India’s population surpass China’s?
10.18 Cohorts. Assume US population of 290 million consists of eight cohorts that
are initially of the same size for eight decades of life. Then modify the cohorts
with time using a death rate of 1%/decade for the first five decades, 20%
for the sixth decade, 25% for the seventh decade, and 100% for the eighth
P1: GFZ
Bibliography 277
decade. One-half the population consists of females who give birth to two
children in their third decade (their twenties) and 0.2 child in their fourth
decade. What will US population be in 50 years in 10-year iterations, cohort
by cohort?
10.19 Combined cycle gas turbine. (a) What is the efficiency of a 500-MWe CCGT
with a 38% efficient Brayton cycle, followed by a 32% efficient Rankine cycle?
(b) How much energy does it take to make a kilowatthour? What is the rate
of consumption of natural gas? What is heat rejection rate for each cycle? (c)
How might water injection increase efficiency? Develop a numerical scenario
that addresses this.
10.20 Carbon-saving CCGT. (a) What is annual reduction in CO2 output if US
switched 200 GWe from coal at 40% efficiency (0.26 kg C/kWh) to CCGTs at
60% efficiency (0.10 kg C/kWh)? Compare results with Eq. 10.31. (b) Show
that the kg C/kWh figures are reasonable.
10.21 Enhanced CCGT. Derive the efficiency for a three-stage CCGT with a second
Rankine system attached at the end?
10.22 Rankine cycle. (a) What is η of a Rankine cycle that requires 0.51 MJ/kg
to convert water to high pressure steam, while the condenser removes
0.31 MJ/kg. (b) An additional 0.16 MJ/kg is used to superheat steam from
335◦ C to 538◦ C (1000◦ F). The condenser now removes 0.38 MJ/kg. What is
the efficiency of the superheat cycle?
Bibliography
Alpert, S. and M. Gluckman (1986). Coal gasification systems for power generation, Ann.
Alpert, S. (1991). Clean coal technology, Ann. Rev. Energy Environ. 16, 1–23.
Benka, S. (Ed.) (2002). The energy challenge, Phys. Today 55(4), 38–75.
Borowitz, S. (1999). Farewell To Fossil Fuels: Reviewing America’s Energy Policy, Plenum Press,
New York.
Cohen, J. (1995). How Many People Can the Earth Support? Norton, New York.
Energy Information Agency (2006). Annual Energy Outlook, Annual Energy Review, Inter-
national Energy, Largest oil and gas fields (1993), Washington, DC.
Fay, J. (1980). Risks of LNG and LPG, Ann. Rev. Energy Environ. 5, 89–105.
Howes, R. and A. Fainberg (Eds.) (1991). The Energy Sourcebook, American Institute of Physics
Press, New York.
Hubbert, K. (1981). The world’s evolving energy system, Am. J. Phys. 49, 1007–1029.
Krenz, J. (1984). Energy: Conversion and Utilization, Allyn and Bacon, Boston, MA.
Matos, G. and L. Wagner (1998). Consumption of materials in the United States, Ann. Rev.
National Research Council (1995). Coal Energy for the Future, National Academy Press,
Washington, DC.
———(1999). Our Common Journey: A Transition Toward Sustainability, National Academy
———(2001). Growing Populations, Changing Landscapes: Studies from India, China, and the
United States, National Academy Press, Washington, DC.
P1: GFZ
Office of Technology Assessment (1991). US Oil Vulnerability, OTA, Washington, DC.

———(1995). Renewing Our Energy Future, OTA, Washington, DC.
Rogner, H. (1997). An assessment of world hydrocarbon resources, Ann. Rev. Energy Environ.
22, 217–262.
Scientific American Energy Panel (1998). Preventing the next oil crunch, Sci. Am. 278(3),
77–95.
US Energy Information Administration (2000). US Crude Oil, Natural Gas and Natural Gas
Liquid Reserves, EIA, Washington, DC.
———(2004). Annual Energy Outlook, EIA, Washington, DC.
———(2004). International Energy Outlook, EIA, Washington, DC.
———(2005). Monthly Energy Review, EIA, Washington, DC.
Weisz, P. (2004). Basic choices and constraints on long-term energy supplies, Phys. Today
57(7), 47–52.
Yergin, D. (1991). The Prize: The Epic Quest for Oil, Money and Power, Simon and Schuster,
New York.
P1: GFZ
11
Energy in Buildings
“Heat goes by itself from the hotter body to the colder body.”
Energy use in buildings accounts for 39% of total energy use. Reducing this frac-
tion could significantly stabilize national security, improve the environment and
enhance the national economy. Residential and commercial buildings consumed
energy totaling 37.5 quads/year in 2001, at a cost of $300 billion/year. Buildings
dominate the use of electricity at 26 quads/year, which is 67% of US electricity con-
sumption of 40 quads/year. On a household basis this amounts to $1391/year ($915
electricity, $476 for heating). The energy use trend in buildings is a success story.
Buildings built prior to the oil embargo of 1973–74 were often an energy disaster,
built without insulation, wasting winter heat and summer air conditioning alike.
Buildings now consume one-half of their former level as energy intensiveness of
big buildings dropped from 270,000 to 100,000 Btu/ft2 year of primary energy.1 But
these gains are being countered by homes that have grown from 1400 ft2 in 1970 to
today’s 2225 ft2 due to more bathrooms and other design extras. In 2000, US home
ownership stood at 66%.
Summer heat gains and winter heat losses are caused by temperature differ-
ences through walls, ceilings, and floors (50–70%), and windows (15–25%), and
by air infiltration (20–30%). Before the oil embargo, leaders of the expansive and
burgeoning building industry were not engaged in coordinated energy research.
The embargo changed this, catalyzing serious research and development at the
Lawrence Berkeley Laboratory and Princeton University on the use of energy in
buildings. The State of California established the first energy standards for build-
ings in 1975. Many states and the federal government followed California’s lead.
One quick way to find out if a house has major heat leaks is to set its thermostat
to the maximum level on a cold day, say 0◦ C (32◦ F). If after an hour the furnace
cannot raise the temperature over 30◦ C (86◦ F), it is likely the house has major heat
leaks or the furnace needs repair. More accurate methods of detecting heat leaks use
1
US home heating modes in (2000/1980/1960/1940) in % of use: natural gas (51/53/43/11),
electricity (30/18/2/0), oil (9/18/32/10), bottled gas (7/6/5/0), coal (0.1/0.6/12/55), wood
(2/3/4/23), solar (0.04/0/0/0) [US Census Bureau].
279
P1: GFZ
280 11. Energy in Buildings
infrared scanners to spot the leaks and blower doors to discover infiltration paths.
“House doctors” can perform energy audits to recommend additional insulation,
leak-caulking, clock thermostats, water-saver shower heads, pipe insulation wrap,
insulation for water heaters, automatic timers for lights, compact fluorescent lamps,
improved appliances, passive solar features, and so forth.
11.1 Heat Transfer

The easiest approach to reducing heat loss is to reduce the thermal conductivity of
walls, ceilings, floors, and windows. Heat flow through a slab is
d Q/dt = (k A/L)T, (11.1)
where k is thermal conductivity, A is slab area, L is thickness, and T is temper-
ature difference between the slab’s opposite surfaces. Writing dQ/dt as Qdot, the
temperature difference across a slab is
T = Q̇(L/k A), (11.2)
which is identical in form to Ohm’s law (V = IRohm ). Temperature difference T
is analogous to voltage difference V, heat current Qdot is analogous to electric
current I = dq/dt, and thermal resistance L/kA is analogous to electrical resistance
Rohm . Thermal conductivity transmits heat from a higher temperature to a lower
temperature, while electrical conductivity transmits current from a higher voltage
to a lower voltage and converts electric power to I 2 R thermal power.
Thermal resistance, or the R-factor, is defined as R = L/kA for a standard area of A
= 1 m2 (1 ft2 ). A builder purchasing insulation may specify 100 ft2 of R13 insulation
for 2 × 4 stud walls, but an electrician does not specify area when buying a 100-
ohm resistor. The thermal transmittance U-factor, is defined as the inverse of thermal
resistance R, or U = 1/R.
Heat flow through a slab of material is thus given by
Q̇ = Q dot = AT/R = U AT. (11.3)
SI units are used in 185 nations, while English units are used in the United States.
SI and English units are as follows: Heat flow, watts (Btu/hour); area, m2 (ft2 ), T,
◦
C (◦ F); R, m2 ◦ C/W (ft2 ◦ F h/Btu) and U, W/m2 ◦ C (Btu/h ft2 ◦ F). R19 walls and
U1 windows in the United States become R3.4 and U6 elsewhere. The conversion
factors between the two cultures are (Table 11.1)
RSI = REnglish /5.67 (11.4)
USI = 5.67 × UEnglish . (11.5)
11.1.1 Series Heat Paths

The R-value of several materials layered in series is the sum of the R-values for the
individual layers. For example, the composite R-value for a wall with two layers
P1: GFZ
11.1. Heat Transfer 281
Table 11.1. R-Values of building materials

RSI REng
Wood siding shingles 0.15 0.9
Stucco, 1 cm 0.014 0.1
Plywood, 1 cm 0.09 0.5
Softwood, 1 cm 0.09 0.5
Concrete block, 30 cm (12 ) 0.2 1.3
Brick/cm 0.014 0.1
Gypsumboard, 1 cm 0.06 0.3
Hardwood floor, 1 cm 0.06 0.3
Carpet 0.11 0.6
Asphalt roof shingles 0.08 0.5
Wood roof shingles 0.17 1.0
Insulation, 10 cm (4 ) 2.3 13
Insulation, 15 cm (6 ) 3.3 19
Polystyrene, 1 cm 0.35 2.0
Polyurethane, 1 cm 0.44 2.5
Gas-filled panels, 2.5 cm (1 ) 2.5 14
Gas-filled panels, 10 cm (4 ) 5–10 28–56
Straw bale, 40 cm (16 ) 6–9 33–50
Glass, 3 mm (0.12 ) 0.005 0.03
Convection 0.04–0.02 0.2–1
Radiation 0.2ε 1ε
Convection depends on wind conditions, drapery

and temperature. Radiation depends on emissivity
ε and temperature.
of plasterboard and a layer of insulation is
Rtotal = R1 + R2 + R3 = L 1 /k1 + L 2 /k2 + L 3 /k3 = 1/Utotal , (11.6)
which gives a loss rate, dQ/dt = Utotal AT.
11.1.2 Parallel Heat Paths

Buildings have many inside-to-outside, parallel heat paths through walls, win-
dows, roofs, floors, and infiltration. As with parallel electric circuits, total heat
flow is the sum of the parallel heat flows, determined from the individual UA
values,
d Q/dt total = d Q/dt 1 + d Q/dt 2 + d Q/dt 3 = (U1 A1 + U2 A2 + U3 A3 )T. (11.7)
Two electrical resistors in parallel have an effective resistance of Reff =

R1 R2 /(R1 + R2 ). If R1 is much larger than R2 , the effective resistance is determined
from the smaller value, or Reff ≈ R2 . If insulation paths are closed with very large
R-values, a point of diminishing returns is reached since infiltration will dominate
with Reff = R2 = Rinfiltration .
P1: GFZ
11.1.3 U-Factors for Radiation and Convection

Surfaces lose heat by convection and radiation in parallel according to
d Q/dt = d Q/dt radiation + d Q/dt convection (11.8)

d Q/dt = σ A(εi Ti4 − εo To ) + h A(T)
4 5/4
(11.9)
where σ is the Stefan–Boltzman constant, A is surface area, ε is emissivity, and h

is a convection constant that depends on geometry (h upward > h vertical > h downward ),
air flow and type of surface. Convection and radiation are not directly analogous
in circuit theory since they do not depend on T to the first power, but they can
be simplified to the form dQ/dt = UAT for easier calculations.
11.1.4 Convection
Convection can be rewritten as
d Q/dt conv = h A(T)5/4 = (hT 1/4 )AT = Uconv AT. (11.10)
The parameter Uconv = hT 1/4 is relatively constant since T 1/4 varies slowly. Its
inverse varies from Rconv-SI = 0.04–0.2 (Rconv-Eng = 0.2–1) with larger values outside
in the wind and smaller values inside buildings.
11.1.5 Radiation
Net radiation flow from a surface at temperature T1 located in outside ambient
temperature T2 is
d Q/dt = σ Aε(T1 4 − T2 4 ) (11.11)
with emissivities assumed equal for simplicity, ε1 = ε2 = ε. Since temperature dif-

ference is much less than ambient absolute temperature (T ≈ 30 K T1 or T2 ),
we can linearize net radiation flow to
d Q/dt net = (Aεσ T 4 ) ≈ (4εσ T 3 )AT. (11.12)
This gives the net radiation U-factor
Urad = (d Q/dt)/AT = 4εσ T1 3 . (11.13)
Urad at room temperature T1 = 20◦ C (293 K) is
Urad-SI = 4εσ T1 3 = (4ε)(5.7 × 10−8 )(293 K)3 = ε(5.7). (11.14)
A blackbody with ε = 0.9 gives Rrad-SI = 0.2 (Rrad-Eng = 1), which is similar to Rconv .
A stove in a building loses heat first by convection and radiation to the inside wall
surfaces, then by conduction through the walls, windows and infiltration, followed
by convection and radiation away from the outside wall surfaces.
P1: GFZ
11.2. Single- and Double-Glazed Windows 283
11.2 Single- and Double-Glazed Windows

The use of passive solar energy in buildings (Chapter 12) depends on glass that
transmits sunlight readily, but strongly absorbs inside infrared. The predominant
IR wavelength from objects at 300 K is 10 μ, obtained with Wien’s law using the
known properties of the predominant solar wavelength of 0.5 μ at 6000 K, that is
λir = λsun (Tsun /Troom ) ≈ 0.5 μ(6000 K/300 K) ≈ 10 μ. (11.15)
The convection and radiation paths are parallel, giving an effective resistance of
Rrad Rconv
Rpar = . (11.16)
Rrad + Rconv
Convection resistance is larger on inside surfaces as compared to outside sur-
faces. Ignoring this difference, the total R-value for one pane of glass is the sum of
two equal parallel resistances and the conductive resistance of a thin sheet of glass,
or
Rtotal = 2Rpar + Rcond . (11.17)
Using approximate SI units, Rconv = Rrad = 0.2, Rpar = 0.1 and Rcond = 0.005 for
glass,
Rtotal = 2(0.1) + 0.005 = 0.205, (11.18)
in fair agreement with Table 11.2 with RSI = 0.16 (REng = 0.9) for single-glaze. The
conduction R-value for a pane of glass can be ignored since it is much less than the
convection and radiation R-values. With equal resistance on either side of glass, the
glass temperature is halfway between inside and outside temperatures. This can
be seen in a common energy-saving practice. A cold room at 60◦ F (16◦ C) on a cold
day of 0◦ F (−18◦ C) develops ice on the inside of its single-glazed window since the
median temperature is 30◦ F (−1◦ C). The ice won’t form if transparent plastic wrap
is taped over the window frame, giving a cheap double-glaze window.
Double-glaze R-values can be estimated from the wiring diagram in Fig. 11.1.
We approximate the four parallel resistances to be the same. It is not true that they
Table 11.2. U- and R-factors for aluminum frame windows

Window U-Factors USI RSI UEng REng
Single glaze 6.1 0.16 1.07 0.93
Double glaze, 1/2 air space 3.5 0.29 0.62 1.6
1/2 air space, ε = 0.1 2.8 0.36 0.49 2.0
1/2 argon space, ε = 0.1 2.6 0.38 0.46 2.2
Triple glaze, 1/2 argon space, ε = 0.1 2.0 0.5 0.36 2.8
Quad glaze, 1/4 krypton, ε = 0.1 on 2 panes 1.2 0.83 0.22 4.7
A factor of 5 reduction can be obtained with more sophisticated windows,

but solar gain may be reduced (American Society of Heating, Refrigeration,
and Air Conditioning Engineers, 1993).
P1: GFZ
Figure 11.1. Double-glaze heat transfer circuit diagram.
are the same, since the still air between two sheets of glass makes convection less
effective in that region, and outside convection is greater than inside convection.
Ignoring the R-value of glass, we obtain Rtotal-SI = 4 × 0.1 = 0.4 (Rtotal-Eng 2.2),
which is fairly close to the measured value of Rtotal-SI = 0.3 (Rtotal-Eng 1.6). Our theory
doubles the R-value with the second pane of glass, but a figure that is too optimistic
since the ratio is actually 1.7. Further savings are possible when low-emissivity
(low-E) coatings are used to reflect (rather than absorb) IR back into the room
(Section 14.3). In addition pulled drapes reduce convection and radiation losses,
particularly in the night when outside temperature is lowest. In general, windows
with lower U-factors have lower solar heat gain and lower visible transmission.2
11.3 Annual Heat Loss

As defined previously, the rate of heat loss through a surface area A is
d Q/dt = U AT. (11.19)
Annual heat loss is obtained by rearranging this differential equation into a
product of heat loss rate dQ/dt and time spent at that rate dt, or
d Q = U ATdt. (11.20)
Rather than use infinitesimal dt for the time interval, we use the hourly interval
T in a finite sum of hourly inside and outside temperature differences over an en-
tire year. Heating severity is proportional to the product Tt. Since only heating
is of concern here, retain just the terms that have Toutside colder than Tbase (a defined
base temperature). This sum gives the quantity degree-hours per year (dh/yr):

8760
dh/yr = (Tbase − Toutside )i (1 h). (11.21)
i=1
2
For example, two different double-glaze windows had these characteristics:
A: U-factor (0.49), solar heat gain coefficient (0.62), visible transmittance (0.64)
B: U-factor (0.37), solar heat gain coefficient (0.33), visible transmittance (0.54).
P1: GFZ
11.3. Annual Heat Loss 285
There are 8760 h in a year (without leap year) and t is a 1-h increment. Toutside is
the average outdoor temperature at a given hour and Tbase is the base temperature
defined as 65◦ F which takes into account 3◦ F free temperature from the internal heat
of people and machines (Section 11.5). A division by 24 converts degree hours to
degree days. We give the heating dd summation:

8760
dd/yr = (65◦ F − Toutside )i (1 h)/24, (11.22)
i=1
for Toutside < 65◦ F. See Fig. 11.2 for a map of US degree days and Table 11.3 for data
from selected cities. We can now determine the annual heat loss over a single path j:
Qj = Uj Aj (dd/yr)(24 h/day). (11.23)
Since a building has many parallel loss paths through its envelope, total annual
heat loss is a summation over n paths,

n
Qtotal = (dd/yr)(24 h/day) Uj Aj . (11.24)
j=1
The total heat loss must be increased by about 25% to account for infiltration
losses. Annual fuel consumption is obtained by dividing total heat loss by furnace
efficiency η (70–90%), which should be reduced to take into account heat duct
losses of perhaps 20%. Daily heat loss is determined from the number of heating
degree-days on a particular day.
Superinsulated houses with 20◦ F (10◦ C, or more) free temperature from 1–2 kW
internal heat can greatly reduce losses (Section 11.6). Passive solar can give further
reductions (Chapter 12). A word of caution: the concept of heating degree days is not
useful for warmer climates because it ignores daytime storage of energy that is used
in the early evening. Degree-day calculations also ignore the time-dependence of
the effects of thermal inertia, the thermal flywheel, from passive solar energy (Section
12.6).
In a similar fashion, cooling degree-days per year (cdd) are defined for air condi-
tioning, but with T = Toutside − Tcool , with Tcool = 75◦ F (24◦ C). The annual energy
needed for air conditioning is
Q/year = (U A)(cdd/year)(24 h/day)/COP (11.25)
in watt hour/year or Btu/year. The coefficient of performance COP of an air condi-
tioner is the ratio of heat removed from the building (Qin ) to the work required to
remove it (W); typical COP = Qin /W = 2–3. This accounting is misleading since
it ignores the 30–60% efficiency of power plants. Thus, the total energy consumed
to remove Qin with power plant efficiency η = 33% is the same as the heat energy
removed (Section. 14.4):
E total = Qin /(COP)(η) = Qin /(3)(1/3) ≈ Qin . (11.26)
However a combined cycle gas turbine with η = 60% gives Qin /(3)(0.60) ≈ 0.56Qin .
P1: GFZ
Figure 11.2. Heating degree days with a 65◦ F baseline.

P1: GFZ
11.3. Annual Heat Loss 287
Table 11.3. Heating and Cooling

Degree-Days in ◦ F-days/year
City Heating Cooling
Atlanta, GA 3095 1589
Boston, MA 5621 661
Chicago, IL 6200 900
Fresno, CA 2650 1671
Los Angeles, CA 1819 600–1100
New York, NY 4900 1000
St. Cloud, MN 8868 426
Salt Lake City, UT 5983 927
Seattle, WA 4727 183
Tucson, AZ 1752 2814
Washington, DC 4211 1415
11.3.1 Window Losses

Considerable energy is lost through windows. The average living unit has a floor
area of 140 m2 (1500 ft2 ) and window area of 20 m2 (15% of floor area, 225 ft2 ) with
an SI U-factor of 4 (English 0.7). The annual window energy loss for a 2800◦ C day
(5000◦ F day) heating season in SI units is
d Q/dt = (U A)(24)(dd/yr) = (4 W/m2 ◦ C)(20 m2 )(24 h/day)(2800 ◦ C day)

= 1.9 × 1010 J/year. (11.27)
In practical units this is 18 MBtu/year, or 3.2 bbl/year of oil (equivalent). If

furnace/duct efficiency is 2/3, the total fuel used for windows is (3.2 bbl/year)(3/2)
∼
= 5 bbl/year. For 100 million US residences the loss of energy through windows is
500 Mbbl/year = 1.3 Mbbl/day, or 3% of total US energy use. A household savings
of $300/year could be saved if Boston homeowners switched from single glaze to
triple glaze.
11.3.2 Infiltration Losses

About one-fourth of heat is lost from cold air infiltrating into a house through holes
and small cracks. Heat transfer from infiltrating air is given by
d Q/dt infil = (dm/dt)cT, (11.28)
where dm/dt is the infiltration rate of air mass. The rate of exchanging air in build-
ings is described in terms of air changes per hour, or ach. A building with an
infiltration rate RACH = 1 ach turns over 100% of its air in 1 h. Rewriting the rate of
heat loss from infiltration gives
d Q/dt infil = (Vρ)RACH cT, (11.29)

P1: GFZ
where Vρ is the mass of the air in the house (volume × density) and c is the specific
heat of air. The annual heat loss over the year is
d Q/dt infil = (Vρ)RACH c(dd/yr)(24 h/day). (11.30)
Annual infiltration loss is estimated using these assumptions:
r house volume V = 2.5 m × 140 m2 (8.2 ft × 1500 ft2 )
r infiltration causes RACH = 1 ach
r air density ρ = 1.3 kg/m3 (0.0735 lb/ft3 )
r air specific heat c = 1004 J/kg ◦ C (0.24 Btu/lb ◦ F)
r average 2800◦ C day/year (5000◦ F day/year).
This gives
d Q/dt = (140 × 2.5 m3 )(1 ach)(1.3 kg/m3 )(1004 J/kg ◦ C)(24 h/day)
× (2800 ◦ C day/year)
= 3 × 1010 J = 30 MBtu/year = 5.5 bbl/year. (11.31)
Infiltration costs the average homeowner whose furnace/duct efficiency η = 2/3
about (5.5 bbl/year)/(2/3) = 8 bbl/year. For a nation of 100 million housing units,
this amounts to 800 Mbbl/year = 2 Mbbl/day. If all houses were tightened by 50%
to 0.5 ach, the nation would save roughly 1 Mbbl/day, but the level of radon and
pollution would be increased (Section 7.7). Air-to-air heat exchangers are used in
superinsulated houses to retain 75% of the heat in exiting air by heat transferring
it to incoming air.
11.4 Energy Standards

California energy standards for buildings and other areas have made a dramatic
difference since the oil embargo. Fig. 11.3 shows that per capita electricity use in
California remained approximately constant from 1976 to 2000 at 7500 kWh (aver-
age power = 0.86 kWe ), while the US per capita electricity use rose from 8000 to
12,000 kWh/year (from 0.9 to 1.4 kWe ). California’s 1975 Warren–Alquist legisla-
tion has been very successful, and other states have followed this lead.
The California Energy Commission divides the state into 16 climate zones with
different packages of standards for a particular zone. The regulations examine each
building on its total heat performance, rather than requiring absolute adherence
to individual criteria. It is acceptable for a builder to lose excessive energy in one
area as long as another area makes up for the losses. Common English units are
used for furnaces (Btu/h) and insulation (ft2 ◦ F h/Btu), but not for electrical heating
(watts). Three package choices for 1999 standards in Zone 5 (San Luis Obispo) were
as follows:
r Passive solar: Ceiling (R30), wall (R13), wall with heavy mass density of
40 pounds/ft2 (R2.4), raised floor (R13), slab floor (R7), double glazing (U 0.45),
P1: GFZ
11.4. Energy Standards 289
Figure 11.3. Total electricity use per capita in California and US (1960–2001). California
energy regulations kept per capita electricity use constant, while the US increased by 50%.
[A. Rosenfeld, California Energy Commission, 2002]
south-facing glazing area (more than 6.4% of floor area), non south-facing glazing
(less than 9.6% of floor area).
r Thermos bottle: Tight insulation without passive solar, ceiling (R30), wall (R19),
slab floor (R7), raised floor (R19), double glazing (U0.45), glazing area less than
14%.
r Electrical heating: No natural gas available, ceiling (R38), wall (R25), slab floor
(R7), raised floor (R30), triple glazing (U 0.4), glazing area less than 16%.
The R-values refer to insulation in the walls. Framing affects the composite U-
factor for a wall since 2 × 4 studs every 16 in comprise 10% of wall area. For
R1 13 wall insulation and R2 4.4 for 4 in of depth, the composite U-factor including
framing is
Uinsul/studs = 0.9/R1 + 0.1/R2 = 0.9/13 + 0.1/4.4 = 0.07 (11.32)

Rinsul/studs = 1/U = 1/0.07 = 11. (11.33)
The studs reduce the effective wall R-value from R13 to R11, but this result is
more than countered by the layer of plaster, the outside framing layer and the
convection/radiation R-values. As a compromise, regulations use the R-value of
the insulation by itself, ignoring losses from studs and gains from the ignored series
resistance, as they tend to cancel one another.
P1: GFZ
Table 11.4. Energy losses from three houses

1955 House 1978 House 1999 House
area (ft2 ) U UA U UA U UA
Floor 1600 0.2 320 0.08 128 0.053 85
Ceiling 1600 0.2 320 0.047 75 0.033 53
Windows∗ 400 1.1 440 1.1 282 0.45 101
Walls∗∗ 912 0.3 274 0.08 84 0.053 58
Total UA, loss in Btu/hr-◦ F 1354 569 297
A house built in the 1950s without regulations is compared with houses built under 1980 and 1999
California Energy Commission standards. Relative heat loss UA is often called the lossiness, and it is
given in Btu/hr ◦ F with U in Btu/ft2 ◦ F hr.
∗ Glazing area: The 1955 house has 400 ft2 of windows, the 1978 house has 1600 × 16% = 256 ft2 of
windows, and the 1983 house has 1600 × 14% = 224 ft2 of windows.
∗∗ Wall area: The 1950 house has a total siding (walls plus windows) area obtained by multiplying the
perimeter of the house times the height, or (32 + 32 + 50 + 50)(8) = 1312 ft2 . The net wall area for the
1955 house subtracts 400 ft2 for windows to give 912 ft2 . The 1978 house wall area is 1312 ft2 minus
256 ft2 (windows) for a net wall area of 1056 ft2 . The 1999 house wall area is 1312 ft2 minus 224 ft2
(windows) for a net wall area of 1088 ft2 .
11.4.1 Heat Loss in Three Houses

We now compare energy loss from a house with poor insulation (circa 1955) to
houses that satisfied the 1977–82 standard and houses that satisfied the 1999 ther-
mos bottle standard. To determine the total heat loss design requirements, we
multiply the U-factors and the corresponding areas of wall, ceilings, floors, and
windows. We assume a 1600-ft2 house in moderate 4000◦ F-day weather. The house
has a 32 × 50 rectangular plan and a window area of 400 ft2 . The results are given in
Btu/h ◦ F, called lossiness, which represents the heat lost per degree of temperature
difference.
Table 11.4 shows that the 1950 house with a lossiness of 1354 Btu/h ◦ F consumes
2.4 times that of the 1978 house (569 Btu/h ◦ F) and 4.5 times that for the 1999 house
(297 Btu/h ◦ F). In addition the losses from infiltration and furnace/ducts should
be included. Relative loss rates are easily converted into annual energy losses by
multiplying by 24 times the number of ◦ F-days. Nonstandard housing designs,
such as berm houses (which are located partially underground), are analyzed with
computer codes to determine energy performance. See Figs. 11.4 and 11.5 for energy
conserving measures in buildings.
11.5 Scaling Laws for Buildings

Energy use by large buildings is load dominated since large buildings have consid-
erable internal heat from equipment, people and lighting. Big buildings have less
surface area per unit volume, with internal energy supplying much of the sur-
face losses. This makes large buildings fairly independent of climate while small
P1: GFZ
11.5. Scaling Laws for Buildings 291
Figure 11.4. Potential energy savings in housing. The potential energy savings are figured
in units of 100 million Btu/year, which is 17 bbl of oil per year. Conservation measures
are applied to a northern California house (1200 ft2 , 3000◦ F days) in terms of additional
contractor investment. If all the energy-savings measures are adopted, it would cost $2700
to reduce energy use by a factor of 3 from the case of an uninsulated house, and by a factor
of 2 from an R11 ceiling house. [A. Rosenfeld, Lawrence Berkeley Laboratory]
buildings have energy bills proportional to the heating and cooling degree days
they experience. Energy use in houses is skin dominated since houses must replace
heat losses through their envelopes. The physical difference between large and
small buildings is easily seen through scaling law relations. (Section 1.3 used scal-
ing laws to give approximate critical masses for nuclear weapons.) Scaling laws for
buildings determine the free temperature of buildings as a function of size, as well
as for superinsulated houses.
Heat loss is proportional to a building’s surface area L 2 and the temperature
difference T between inside and outside:
d Q/dt loss = Ueff L 2 T, (11.34)
where Ueff is the effective thermal transmittance (Ueff = 1/Reff ) for the building that
takes into account all energy leaks. On the other hand, internal heat gain is proportional
P1: GFZ
Figure 11.5. Cost-benefits of conversation measures in buildings. (A. Rosenfeld, Lawrence

Berkeley National Laboratory.)
to building volume (floor area times a fixed ceiling height of about 3 m):
d Q/dt gain = G L 3 , (11.35)
where G is internal heat load per unit volume. A typical house has about 1 kW
of free heat (3400 Btu/h), while office buildings have an internal gain flux on a
floor-area basis of f = 66 W/m2 (6 W/ft2 ). The volume gain G is f /H where H is
P1: GFZ
11.5. Scaling Laws for Buildings 293
the height of one floor, obtained from

d Q/dT gain = G L 3 = f nL 2 = f L 3 /H, (11.36)
where the number of floors n = L/H. Because of internal gains, the inside of a
typical unheated house is 2◦ C (3 to 4◦ F) warmer than the outside temperature.
Let’s consider the simple case of walls without mass and specific heat, ignoring
the time dependence of warm-up and cool-down (Section. 12.5). For a building
without a furnace, the gains and losses are equal, and T = Tfree (without a furnace
here):
d Q/dt gain = d Q/dt loss = G L 3 = Ueff L 2 Tfree , (11.37)
or
Tfree = G L/Ueff . (11.38)
Thus, Tfree is proportional to occupation density G and building size L. Large
free temperatures are observed in big buildings (large L), buildings with very good
insulation (small Ueff ) and buildings with large internal heat loads (large G). We
will apply the scaling model to large buildings (large L) and superinsulated houses
with considerable insulation (small Ueff ).
A temperature difference is needed to force heat power through the walls. Such
is the case of elevated temperatures created by placing a blanket over a 100-W light
bulb. The blanket’s high thermal resistance blocks heat flow, raising the temperature
of the interior of the blanket. The higher interior temperature pushes the 100 thermal
watts through the blanket, but it may also start a fire. Free temperature Tfree in
buildings can save considerable energy.
Thermostats do not call for extra heat until Toutside drops Tfree below Tthermostat .
The outside temperature at which the furnace comes on (ignoring time delay from
thermal inertia) is called the balance point of a building. The balance point temper-
ature of a typical building is
Tbalance = Tthermostat − Tfree = 68◦ F − 3◦ F = 65◦ F = 18.3◦ C. (11.39)
This temperature is universally chosen as the base temperature in the degree-day
formula since the furnace turns on when temperature goes below Tbalance . At the
balance point, the internal heat gain without the furnace balances the heat losses. As
building sizes increases, the increased dQ/dtgain raises Tfree = GL/Ueff , lowering
the value of Tbalance at which heat is first needed.
At outside temperatures below the balance point, the net heat loss rate is
d Q/dt net = d Q/dt loss − d Q/dt gain = Ueff L 2 (T − Tfree ) = Ueff L 2 (T − G L/Ueff ),
(11.40)
where T = Tthermostat − Toutside . Note that length L appears in two ways: The
multiplicative Ueff L 2 term, which is the lossiness of the building, increases with con-
ductivity and size. The subtractive GL/Ueff term for free temperature reduces losses
by increasing free temperature and lowering the balance point.
P1: GFZ
Decreased thermal transmittance (Ueff = 1/Reff ) saves energy in two ways.
(1) Lowering the multiplicative Ueff saves energy proportionally since

dQ/dtnet α Ueff L 2 . A superinsulated house can have five times (or more)
thermal resistance than a normal house, raising walls from R10 to R50. If this
is done for all pathways, Usuper = Unormal /5. The multiplicative role of Usuper ,
reduces heating bills to 20% of their former value.
(2) The subtractive free temperature allows houses to be run at effectively lower
temperatures. Let the internal energy of a normal house be 1 kW (3400 Btu/h)
with lossiness Ueff L 2 = 1100 Btu/h ◦ F. Free temperature is obtained from a heat
balance:
1 kW = 3400 Btu/h = Ueff L 2 Tfree = (1100 Btu/h ◦ F)Tfree-normal , (11.41)
giving Tfree-normal = 3◦ F and Tbalance-nomral = 68◦ F − 3◦ F = 65◦ F. A superinsulated

house with 20% of its former lossiness (Ueff L 2 = 220) has a Tfree that is five
times larger than a normal house.
Tfree/super = (Unormal /Usuper )Tfree/normal = 5Tfree/normal = 5 × 3◦ F

= 15◦ F = 8◦ C. (11.42)
Considerable energy saving is seen by examining thedegree day distribution func-

tion for a particular location. For days when Toutside is greater than the balance point
of the building, 100% of the energy is saved. On other days, a fractional energy sav-
ings is realized. For example, if the free temperature of a superinsulated house
is 15◦ F, the balance point is Tbalance = Tthermostat − Tfree = (68◦ F − 15◦ F) = 53◦ F.
The furnace is not engaged on days that are warmer than 53◦ F, which gives 100%
savings. For a just-freezing day of 32◦ F, the furnace supplies 53◦ F − 32◦ F = 21◦ F of
what a typical house needs, 65◦ F − 32◦ F = 33◦ F, which reduces normal heating bills
by 21◦ F/33◦ F = 64%. But, we have not included the multiplicative factor of 20%,
which gives the actual fuel bill reduction to 64%/5 = 13% of its normal value.
Large families with a greater number of occupants that use and produce more
thermal power further enhance savings. Twice the thermal output doubles free tem-
perature from 15◦ F to 30◦ F. Some super-super insulated houses in Saskatchuwan,
Canada, with (10,000 ◦ F days/year) use only $100 of natural gas to get through the
winter. We now see that it is theoretically possible for a house with extremely small
lossiness to function with the heat of “two cats fighting,” but economics makes this
only a pedagogical example.
11.5.1 Clock Thermostats

Clock thermostats regulate house temperature so that house temperature is raised
just before residents awake or return from work. Clock thermostats also can be
used to lower temperatures further than a person would normally do with a reg-
ular thermostat. We estimate savings available with clock thermostats for several
scenarios.
P1: GFZ
Problems 295
Suppose a building with a balance point of 65◦ F (Tthermostat = 68◦ F, Tfree = 3◦ F)

has an outside temperature Toutside = 35◦ F. The furnace burns fuel at a rate
d Q/dt = α(65◦ F − 35◦ F) = 30α, (11.43)
where α is Ueff L divided by furnace efficiency η. The fuel used during a 24-h day
2
with the thermostat at 68◦ F is 30α × 24 h = 720α. If the 68◦ F thermostat temperature
is lowered 1◦ F to 67◦ F, fuel rate drops to 29α, saving (30α − 29α)/30α. = 3% of fuel.
If all thermostats in the United States were lowered 1◦ F, the energy savings would
be (0.03)(5 Mbbl/day) = 0.15 Mbbl/day (equivalent) on an annual basis.
11.5.2 Overnight Setback

The saving from an overnight, 8-h, lowering of thermostat from 68◦ F to 50◦ F is
Q/day = α(68◦ F − 50◦ F)(8 h) = 144α, (11.44)
◦
which is 20% (144α/720α) of the heating bill for a constant setting of 68 F. The extra
heat required to raise the temperature of the interior mass at daybreak is balanced
by raising the temperature 1 h early and lowering it 1 h early at night.
11.5.3 Absent Resident

If a person is home only 6 waking hours a day, the temperature can remain at
50◦ F for 18 h. This saves Q/day = α(68◦ F − 50◦ F)(18 h) = 324α, which is 45%
(324α/720α) of the heating bill.
Problems
11.1 Units. Show RSI = REnglish /5.69 and convert Btu, Btu/h, Btu/h ◦ F,
Btu/h ft2 ◦ F, Mbbl/day, and quad to SI units. (We apologize to the world’s SI
majority.)
11.2 Bulb switch. Six 100-watt incandescent lights are replaced by 40-watt fluo-
rescent tubes that produce five times the lumens/W. (a) How many tubes
are needed to retain the same illumination? (b) What are energy savings for
a year’s full-time operation? What are the air conditioning savings with a
COP of 3? (c) What are coal savings over 10 years with 1 kWh = 0.9 pound
of coal?
11.3 Your household. Estimate your monthly electricity consumption based on
appliance labels and your use-time estimates. Compare this estimate to your
utility bill.
11.4 Walls. (a) Calculate the U-factor for a windowless wall that has an interior
panel of 1/2-in gypsum board, an outside panel of 1/2 in plywood onto
which are nailed wood shingles. Compare this U-factor to the 1999 Califor-
nia Package thermos bottle standards. (b) What is U-factor if radiation and
convection are considered? (c) What is the heat flow rate for T = 30◦ F for
P1: GFZ
120 ft2 of (a) and (b) walls. (d) What is the allowable heat loss using thermos
bottle standards?
11.5 Windowed walls. (a) What is heat flow/◦ F through the wall of problem 11.4
if it holds a 15-ft2 single-glaze window (b) Double glaze, same size.
11.6 Optimal thickness. Thicker insulation saves more energy, but costs more.
Derive an equation for optimal thickness using insulation costs C ($/ft2 in),
R-value (R/inch), climate in dd/yr (◦ F days/year), fuel costs F ($M/Btu) and
furnace/duct efficiency η.
11.7 R and mass location. Separate the thermal resistance R and the thermal capac-
ity C characteristics of matter into pure R and C separate entities (analogous
to R and C in circuits). Do this with a separation of pure R styrofoam insula-
tion (little mass) and pure C for masonry (little thermal resistance). Compare
the situation with R to the outside vs. C to the outside. Compare results to
an electrical circuit.
11.8 House losses. A 30 ft × 50 ft × 8 ft house has 500 ft2 of windows, R13 walls,
R30 ceilings, R3 windows, and R5 floors. (a) What is the heat loss/◦ F and loss
per 8000 ◦ F day/year? (b) Compare to the CA 1999 thermos bottle standard.
11.9 Multi-layered wall. A wall consists of 3 in of mineral fiber (R11), two 1/2-in
gypsum boards (R0.45 each) and two convection/radiation surfaces (R0.4
each). (a) What is the effective R-value of the wall? (b) Draw the wiring
diagram for the wall. (c) Replace the mineral fiber with polyurethane at
R6.3/in and again find the R-value of the wall.
11.10 Condo versus house. What are the energy needs for a house (one floor,
30 ft × 50 ft) and for a center-unit condo (2 floors, 25-ft wide × 30-ft deep)
in 6000 ◦ F day/year. Assume furnace/duct η = 2/3, R30 ceilings, R19 walls,
R10 floor, and no windows for the sake of simplicity.
11.11 Building scaling law. Compare energy loss per unit volume for cubic build-
ings of 50 ft and 100 ft on a side. The window area is 16% of total floor area and
with R30 ceilings, R19 walls, R10 bottom-floor and R1 windows. Determine
loss in Btu/h ◦ F for 6000 ◦ F day/year with furnace/duct η = 2/3.
11.12 Degree days. How many degree days does Wauwatosa, Wisconsin, have in a
winter that has 20◦ F for 3 months, 40◦ F for 2 months, and 55◦ F for 2 months?
11.13 Cold day. How many degree days did Wauwatosa have on a day when it
was −20◦ F from midnight to 6 am, 0◦ F from 6 am to 6 pm, and −10◦ F from
6 pm to midnight?
11.14 Balance point. What are free temperature and balance point for two houses
with lossiness of 1500 and 500 Btu/h ◦ F for internal heat of 3400 Btu/h
(1 kW)?
11.15 Energy saved. How much fuel is used for the two houses of problem 11.14
with η = 2/3 and the degree-day distribution of problem 11.12?
11.16 Heat of day into coolth of night. The 100 ft building of problem 11.10 has
rock thermal storage with specific heat 0.25 Btu/lb ◦ F. (a) If daytime heat of
10 W/ft2 is used to raise storage to 75◦ F, how much cement (water) per ft2 of
floor is needed? (b) How cold will the building be after a 30◦ F 12-h evening,
ignoring infiltration?
P1: GFZ
Bibliography 297
11.17 1 vs. 2 floors. Compare the lossiness of two houses each with total square areas
of 2000 ft2 , one with two floors and the other with one floor. The buildings
have R19 walls, R10 floors, R30 ceilings and R2 windows (14% of floor area).
11.18 Cubic superinsulated. A cubic house is three stories high and 30 ft on a
side. Assume ceiling/floor at R80, walls at R50, and 300 ft2 windows are
triple-glaze, low-emissivity at R6 and infiltration increases lossiness by 25%.
(a) What is its loss coefficient α in dQ/dt = αT? (b) Determine the balance
point Tbalance of the building if the internal heat is 3 kW. (c) How large a
furnace is needed if the outside temperature is −30◦ F? (d) How much fuel
is needed for a degree-day season of problem 11.12 and a 2/3 furnace/duct
efficiency?
11.19 Internal heat. Two cats fighting generate 20 watts (compare to problem 11.20)
of heat in a cubic house, 40 feet on a side, without windows. (a) Equat-
ing energy loss to energy gain, what six-surface R-value is needed to give
30◦ F free temperature? (b) What furnace size (Btu/hour, η = 2/3) is needed
for 68◦ F inside if the outside temperature is 0◦ F? At what rate is natural
gas used?
11.20 Animal scaling laws. The metabolic heat rate for an animal scales with its
mass to 3/4 power (m0.75 ) over 17 orders of magnitude. If an 80-kg man
produces 100 W of heat, what does a 5-kg cat produce? The internal metabolic
heat of animals does not scale with mass to the first power because the fractal-
like fluid-distribution system takes mass. If the lifetime of mammals scales
as m0.2 and their heat-rate scales as m−0.25 , how long do cats live and what is
their heart rate? A 100-g mouse? (see Gillooly et al., 2001).
Bibliography
Adams, E. (Ed.) (2000). Alternate Construction, Wiley, New York.
American Council for Energy Efficient Economy (1984–2004). Studies on Energy Efficiency,
ACEEE, Washington, DC.
American Institute of Architects (1993). Energy Design Handbook, AIA Press, Washington,
DC.
American Society of Heating, Refrigeration, and Air Conditioning Engineers (1993).
ASHRAE Handbook, Atlanta, GA.
Clark, W. (1997). Retrofitability for Energy Conservation, McGraw-Hill, New York.
Daniels, K. (1997). Technology of Ecological Buildings, Birkhauser-Verlag, Boston, MA.
Fisk, W. (2000). Health and productivity gains from better indoor environment and their
relationship with energy efficiency, Ann. Rev. Energy Environ. 25, 537–566.
Gallo, C., M. Sala and A. Sayigh (Eds.) (1998). Architecture: Comfort and Energy, Pergamon,
New York.
Gillooly J., J.H. Brown and G.B. West et al. (2001). Effects of size and temperature on metabolic
rate, Science 293, 2248–2251.
Hafemeister, D., H. Kelly and B. Levi (Eds.) (1985). Energy Sources: Conservation and Renew-
ables, American Institute of Physics Press, New York.
Hunn, B. (1996). Fundamentals of Building Energy Dynamics, MIT Press, Cambridge, MA.
P1: GFZ
Krarti, M. (2000). Energy Audit of Building Systems, CRC Press, Boca Raton, FL.
Kreith, R. and R. West (Eds.) (1997). CRC Handbook of Energy Efficiency, CRC Press, Boca
Raton, FL.
Macriss, R. (1983). Efficiency improvements in space heating by gas and oil, Ann. Rev. Energy
Environ. 8, 247–267.
Meckler, M. (1993). Innovative Energy Designs for the ’90s, Prentice Hall, Englewood Cliffs,
NJ.
Office of Technology Assessment (1992). Building Energy Efficiency, OTA, Washington, DC.
Rosenfeld, A. (1999). The art of energy effeciency: Protecting environment with better tech-
nology, Ann. Rev. Energy Environ. 24, 33–82.
Shurcliff, W. (1988). Air-to-air heat exchangers for houses, Ann. Rev. Energy Environ. 13, 1–22.
Tuluca, A. (1997). Energy Efficient Design and Construction for Commercial Buildings, McGraw-
Hill, New York.
US Energy Information Administration (1999). A Look at Residential Energy Consumption, EIA,
Washington, DC.
P1: GFZ
12
Solar Buildings
“Insulate before you insolate.”

“Glass plus mass prevents you from freezing . . . .”
These slogans contain a great deal of physics. Only thermally tight buildings can be
successfully solar heated since solar flux is a low-density energy source with aver-
age power of 0.2 kW/m2 . Hence, it is good advice to “insulate before you insolate.”
By a quirk of nature, marvelous glass transmits visible light from the sun, while
absorbing infrared radiated from inside buildings. Coupling glass filter with ther-
mal mass gives us low-technology solar heat for buildings and water. Heat capacity
of materials allows us to transfer heat from the hotter body to the colder body with-
out moving parts. On the other hand, high tech solar photovoltaic cells will be a
godsend to the industrial world when they are made economically competitive. PV
cells are already the technology of choice in remote areas. In the meantime, passive
solar is an essentially free energy source for new buildings in warm climates and it
supplies a good boost in cold climates.
12.1 Solar Flux

Solar energy results from a three-step fusion of four protons into a helium nucleus.
The solar flux so is a result of the following phenomena:
r mass of 1 H is 1.0078 AMU and 4 He is 4.0026 AMU
r solar mass is 2 × 1030 kg
r earth-sun distance, the astronomical unit (AU), is 150 million km
r sun will become a giant star at age 1010 years, when 10% of 1 H is consumed.
The mass of four protons is reduced 0.7% when converted to 4 He:
m/m = (4 × 1.0078 − 4.0026)/(4 × 1.0078) = 0.0071. (12.1)
The available solar energy over the sun’s lifetime is
E sun = Msun c2 = (0.0071)(0.1 × 2.0 × 1030 kg)(3 × 108 m/s)2 = 1.3 × 1044 J.
(12.2)
299
P1: GFZ
300 12. Solar Buildings
Figure 12.1. Solar spectrum at

sea level [National Renewable
Energy Laboratory]
The sun’s average power over its 1010 year life is

Psun = E sun /t = (1.3 × 1044 J/1010 year)(1 year/3.2 × 107 s) = 4.1 × 1026 W,
(12.3)
which gives a solar flux above Earth’s atmosphere,
so = Psun /4π(1 AU)2 = (4.1 × 1026 W)/(4π )(1.5 × 1011 m)2 = 1.5 kW/m2 . (12.4)
This value is within 10% of the measured solar “constant” of
so = 1.367 kW/m2 = 434 Btu/ft2 h = 0.13 kW/ft2 = 2.0 cal/min cm2 . (12.5)
The solar flux (Fig. 12.1) at Earth’s surface is reduced by three factors of 2, by
daytime angles, by nighttime darkness and by reflection and absorption by atmo-
sphere and clouds. The first factor of 2 comes from an average of cos θ over a sphere,
where θ is the sun’s angle from the zenith position. The second 2 results from 12 h
of darkness for each average day. The combined factor of 22 = 4 reduction is easily
grasped. The area of Earth’s disk intercepting sunlight is π RE 2 (RE = Earth radius),
but rotation spreads sunlight over 24 h onto Earth’s 4π RE 2 spherical area, which is
four times the disk area. The third factor of 2 is an average of atmospheric absorp-
tion and reflection from atmosphere and clouds. The three factors of 2 reduce solar
flux for the lower 48 states to an average of 1.37 kW/m2 /23 = 0.2 kW/m2 .
12.1.1 Anytime, Anywhere Solar Flux

Solar flux at noon can be estimated from the solar angle θ, which is determined from
a location’s latitude, time and day of the year. Knowledge of θ is used to determine
atmospheric absorption/scattering, and cos θ for horizontal flux. Earth’s spin axis
is tipped 23◦ with respect to the plane of the ecliptic. The tip angle, combined with
latitude angle θL , gives us the solar angle with respect to the zenith at solar noon
P1: GFZ
12.1. Solar Flux 301
Figure 12.2. Solar noon on the equator during sol-

stice and equinox.
(θnoon ) for the two solstice and two equinox days according to the following formula
(Figs. 12.2 and 12.3).
θnoon = θL (spring and fall equinox, March 20 and September 23) (12.6)
θnoon = θL − 23◦ (summer solstice, June 21) (12.7)
◦
θnoon = θL + 23 (winter solstice, December 21) (12.8)
San Diego at 33◦ N has θnoon varying between 10◦ and 56◦ , while Seattle at 47◦ N
has θnoon varying between 24◦ and 70◦ . The value of θnoon for other days is obtained
by fitting a sine function to θnoon values for equinox and solstice days. San Diego
has 14 h of sun in the summer and 10 h in winter. Seattle’s lower sun angle is
countered by 16 h of sun in the summer, but winter is both darker and shorter at
8 h. Everywhere on equinox days the sun rises due east and 12 h later it sets due
west.
The solar flux above the atmosphere so is reduced by absorption and scattering
in air and clouds to s1 . It is further reduced to sh by a factor of cos θ on horizontal
surfaces and to sv by a factor of sin θ on vertical surfaces facing the sun. Referring
to Fig. 12.4, we see how this is so. In this figure, Ao is, for example, an area, which
is orthogonal to solar rays. The solar flux orthogonal to Ao after reduction by atmo-
spheric absorption and scattering is s1 . The horizontal surface that captures all the
Figure 12.3. Solar noon in San Luis

Obispo (35◦ N latitude) during sol-
stice and equinox.
P1: GFZ
Figure 12.4. Horizontal solar flux. Geometrical

reduction of solar flux from s1 to sh = s1 cos θ.
rays passing through Ao is the area Ah . The solar angle θ relates Ah to Ao by

cos θ = Ao /Ah (12.9)
Hence, sh = s1 cos θ , making sh < s1 , except when the sun is in the zenith. Con-
versely, a decrease of sh results from an increase in Ah = Ao /cos θ, spreading the
s1 flux over a larger area. For vertical surfaces facing the sun, such as windows for
passive solar energy, solar flux s1 is reduced to sv by sin θ to sv = s1 sin θs where
sin θ is the ratio Ao /Av .
12.2 Solar Collectors

Solar energy gathered by solar collectors depends on many variables: latitude of
collectors, time of day, season of the year, angle of collectors, and materials and
equipment used for the collectors. These results must be integrated over the day-
time, since passive solar energy is concerned with integrated solar flux and not
instantaneous flux. All of this can make calculations difficult, but our approximate
methods obtain sufficient accuracy to about 10%.
12.2.1 Atmospheric Transmission

We return to the reduction of so to s1 , as air and clouds reflect and absorb sunlight
before it reaches the ground. When the sun is at solar angle θ from the zenith, its
rays pass through more air (proportional1 to the hypotenuse in Fig. 12.5) than when
it is at the zenith position (proportional to the vertical side in Fig. 12.5). The surface
mass density (kg/m2 ) traversed by light is given in units of n Earth air masses, which
increases from n = 1 at a vertical θ = 0◦ to infinity at a horizontal θ = 90◦ (flat Earth
error). The ratio of path lengths in Fig. 12.5 gives the general relation:
n = sec θ = 1/ cos θ. (12.10)
1
Air density falls with elevation h as exp(-h/H), with length of two sides of the triangle
proportional to mass density traversed.
P1: GFZ
12.2. Solar Collectors 303
Figure 12.5. Earth Air Mass. Solar flux passes through more air mass as θ increases.
When the sun is low in the sky at θ = 60◦ , mass density is doubled, as in n =
1/cos 60◦ = 1/0.5 = 2. When the sun is very low at θ = 80◦ , the mass density is
increased by a factor of 6 from n = 1/cos 80◦ = 1/0.17 = 5.7.
A doubling of air mass to n = 2 does not double absorption, as additional air is
actually less effective. The solar flux absorbed, s, in a small amount of mass m
is
s = −λsm, (12.11)
where λ is an absorption constant. This integrates to
s1 = so e −λm , (12.12)
where so is the initial solar flux. The integral mass density of air traversed by
sunlight increases with θ according to
m = nmo = mo sec θ (12.13)
where mo is the air mass traversed at θ = 0◦ . This allows us to write a general
expression for the solar flux at angle θ from the zenith position:
s1 = so exp(−λmo sec θ) (12.14)
The value of λmo is determined from the flux above the atmosphere (so = 1367
W/m2 ) and the maximum flux at Earth’s surface (s1 = 970 W/m2 ) when the sun is
in the zenith, giving
s1 = 970 W/m2 = 1367 W/m2 exp(−λmo ). (12.15)
This gives λmo = 0.34, which gives solar flux at sea level as a function of θ,
s1 = so e −0.34sθ = so e −1/3 cos θ . (12.16)
This can be corrected for site elevation, which depends on the variable density of
air as a function of elevation (problem 12.5).
P1: GFZ
12.2.2 Angle of Collector

Horizontal collector flux is reduced from s1 by cos θ,
sh = (so cos θ )e −1/3 cos θ . (12.17)
Passive solar energy in winter relies on vertical, south-facing glass, which has a
flux at solar noon of
sv = (so sin θ)e −1/3 cos θ . (12.18)
Collectors are typically raised toward the south by an angle φ above the hori-
zontal, giving a flux at solar noon of
snoon = [so cos(θ − φ)]e −1/3 cos θ . (12.19)
12.2.3 Sun at θ = 0◦ and 60◦

When the sun is in the zenith position (θ = 0◦ ), solar flux is reduced to
szenith = so e −1/3 cos θ = so e −1/3 cos 0 = 0.72so . (12.20)
2
Thus, so is reduced by 28%, from 434 to 312 Btu/ft h (for architects) and from
1.37 to 0.98 kW/m2 (for electrical engineers). Sun rays at θ = 60◦ pass through 2
atmospheres, giving
s1 = so e −1/3 cos 60 = so e −1/1.5 = 0.51(434 Btu/ft2 h) = 220 Btu/ft2 h. (12.21)
The solar flux incident on horizontal collectors at solar noon is further reduced by
cos 60◦ = 0.5 to sH = 220/2 = 110 Btu/ft2 h. Note that flux on south-facing vertical
windows at solar noon is not reduced nearly as much as on the horizontal, as sV
= (0.51so )(sin 60◦ ) = 220 × 0.87 = 191 Btu/ft2 h. The resilience of vertical flux in
winter is a tremendous aid to passive solar heating. When snow is on the horizontal
ground, the solarium with vertical glass can be warm. At noon on December 21 in
Minneapolis (45◦ N, θ = 68◦ ), the horizontal flux is reduced to 66 Btu/ft2 h while
south-facing windows receive a respectable 163 Btu/ft2 h on a cloudless day.
12.3 Integrated Solar Flux

Integrated solar flux over daytime hours is relevant for solar heated water or build-
ings. The thermal mass of water and cement respond slowly, as they are not re-
sponsive to instantaneous flux. The empirical data shown in Figs. 12.6 and 12.7 are
similar in shape to the first half-cycle of a sine function. Therefore, we approximate
the direct solar flux as a sine function, ignoring the indirect, diffuse component,
scattered from the entire sky:
sdirect = snoon sin(2π t/T). (12.22)
The maximum amplitude snoon is the flux on the south-facing collector at solar
noon with sunrise at t = 0 and sunset at t = T/2. The empirical direct and scattered
P1: GFZ
12.3. Integrated Solar Flux 305
Figure 12.6. Direct plus scattered horizontal solar flux. Flux is given as a function of time on
solstice and equinox days at 30◦ and 45◦ (north or south) latitude. Flux is given in kW/m2
for solar electricity and Btu/ft2 -h for solar buildings. The values of sH are consistent with
the direct flux calculated in Table 12.1 (Meinel and Meinel, 1977). [Reprinted by permission,
Pearson Education]
Figure 12.7. Direct plus scattered vertical solar flux. Winter flux is greater than summer flux
for passive solar, south-facing windows in the northern hemisphere (Meinel and Meinel,
1977). [Reprinted by permission, Pearson Education]
P1: GFZ
Table 12.1. Integrated Solar Flux

Horizontal collector Raised at θL + 15◦
Latitude Season Sh-noon T/2 I Stip-noon T/2 I

30◦ summer 315 14 2800 245 14 2180
winter 150 10 970 245 10 1550
equinox 255 12 1950 285 12 2180
45◦ summer 285 16 2880 245 16 2450
winter 70 8 360 180 8 930
equinox 195 12 1460 265 12 2010
Solar flux at noon is given in Btu/ft2 h and the number of hours of sunlight
is T/2. The integrated flux is I = sh-noon T/π for horizontal collectors and
I = stip-noon T/π for collectors raised to the south by an angle of the lat-
itude plus 15◦ . The horizontal flux values sh-noon are in agreement with
those given in Fig. 12.6 for direct plus scattered flux.
flux on horizontal surfaces is shown in Fig. 12.6. Note that the equinox daylight is
T/2 = 12 h (T = 24 h) at all latitudes. The summer day is T/2 = 14 h at 30◦ (north
or south) latitude and 16 h at 45◦ latitude. The winter day is T/2 = 10 h at 30◦ and
8 h at 45◦ . Figure 12.7 displays the same direct plus scattered flux, but for vertical
surfaces. Note that vertical surfaces have a much higher flux in winter as compared
to summer, due to the increased value of sin θ.
We determine the horizontal integrated solar flux at equinox in New Orleans
located at 30◦ N with T/2 = 12 h and sh−noon = 255 Btu/ft2 h. Integrating over the
solar day gives an integrated solar flux
T/2 T/2
I = sh dt = sh-noon sin(2π t/T) dt = sh-noon T/π = 255 × 24/π
0 0
= 1950 Btu/ft2 day. (12.23)
The equinox value is almost equally bracketed by the summer solstice value of
2800 Btu/ft2 day and the winter solstice value of 970 Btu/ft2 day. Table 12.1 com-
pares horizontal collectors at 30◦ N and 45◦ N latitude to collectors raised southward
by 15◦ plus the latitude to favor winter when the heat is needed.
Table 12.1 shows that horizontal collectors give very wide variations of integrated
flux during a year, as the daily flux at 30◦ N latitude varies by a factor of 2.9 during
the year while the collector at 45◦ varies by a factor of 8.0. The raised collectors
have a much smaller variation with a factor of 1.4 at 30◦ and 2.6 at 45◦ . But this can
be misleading if seasonal clouds are significant. The average daily integrated flux
is about one-fourth the sum of the daily fluxes of two solstices and two equinoxes.
Since solar space heating is needed only in the winter (Section 12.5), the winter
season is favored by raising the collector toward the south above the horizontal
by 15◦ plus the local latitude. This angle is close to the extreme value of 23◦ plus
the latitude (θ at solar noon on December 23). On the other hand, solar hot water
is needed during the entire year so the collectors are raised at the latitude angle,
P1: GFZ
12.4. Solar Hot Water 307
Table 12.2. Maximum possible annual solar energy yields

Configuration Latitude 103 kWh/m2 106 kJ/m2 kBtu/ft2
Fully tracking (D) 45◦ 2.85 10.3 910
30◦ 3.11 11.3 985
Horizontal plate (D) 45◦ 1.64 5.9 520
30◦ 1.92 6.9 610
Horizontal plate (D + S) 45◦ 1.92 6.9 610
30◦ 2.26 8.1 710
Fixed (+15◦ )(D + S) 45◦ 2.03 7.3 642
30◦ 2.25 8.1 713
D is direct sun rays and S is diffuse sun rays scattered from air and clouds (Meinel
and Meinel, 1977)
which takes advantage of solar oscillation to either side of the equinox position.
However, in the winter only the collectors are usually needed at an angle of 15◦
plus latitude. This is the angle that is usually used for stationary collectors because
(1) winter days are shorter, (2) the winter sun is weaker, and (3) winter feed water
is colder.
Table 12.2 shows data for a variety of collectors at 30◦ and 45◦ N latitude. Col-
lectors that track the sun obtain 25–50% more total energy over the year than fixed
collectors. However, the added value has to exceed the additional cost of solar
tracking, a feature that constrains tracking features to use in more valuable solar
electricity tracking. On the other hand, solar collectors can be manually adjusted
on a seasonal basis to obtain 10% extra energy at a lesser cost. See Fig. 12.8 for a
U.S. map of integrated solar flux.
12.4 Solar Hot Water

Hot water collectors consist of one or two sheets of glass on top of an insulated box
containing an array of copper pipes mounted to aluminum fins. The aluminum
fins are black to absorb the sunrays. Much better than black paint is the selective
absorber, black chrome, which is a good absorber in the visible range and a poor
emitter in the infrared. Inexpensive units use a thermosiphon to move the hot
water, but electric pumps are usually used. The daily rate of energy absorption per
unit area is given by
Qgain /A = I τ α, (12.24)
where I is daily integrated flux, τ is glass transmission, and α is collector absorption.
The energy lost per day by the collector is proportional to the temperature difference
T between metal plates (Tp ) and ambient temperature (Ta ), and the number of
hours per day (t) that the collector is hot. It is given by
Qloss /A = UC (Tp − Ta )t, (12.25)
where U is the collector’s thermal transmittance, and t is 8–16 h. The net available
P1: GFZ
Figure 12.8. Average Annual Solar Energy Resource (1 kWh/m2 = 313 Btu/ft2 ) (DOE).
energy is the difference of gains and losses, that is

Qnet /A = I τ α − UC (Tp − Ta )t. (12.26)
The collector efficiency η is the ratio of the incident solar energy to the useable
energy,
η = Qnet /AI = τ α − UC (Tp − Ta )t/I. (12.27)
P1: GFZ
12.5. Active Solar Space Heat 309
On cooler days, double-glaze collectors have a lower efficiency than single-glaze

since transmission through two sheets of glass absorbs more solar flux. However,
for hot days with high T, double-glaze collectors are more efficient since heat
losses are reduced with a lower UC . A good solar collector uses black chrome with
95% absorption and 10% IR-emissivity, giving system efficiency of η = 50%.
12.4.1 Size of Collectors

The optimal size for hot water collectors varies depending on usage, location, and
water temperature desired. We assume the following to determine collector area
and storage volume:
r 20 gal of hot water/person-day with no reserve for cloudy days
r family of four people
r average flux I = 2000 Btu/ft2 -day and η = 50%
r 120◦ F hot water and 60◦ F cold water.
The energy needed to heat this amount of hot water is
Q/day = 4(20 gal/day)(8.3 lb/gal)(1 Btu/lb ◦ F)(120◦ F − 60◦ F) = 4.0 × 104 Btu/day.
(12.28)
If 50% of the 2000 Btu/ft2 day insolation is useable, the collector area is
A = (4 × 104 Btu/day)/(1000 Btu/ft2 day) = 40 ft2 . (12.29)

Our example system uses 80 gallons storage, for a storage ratio of 2 gal/ft2 . A
40 ft2 system produces more hot water than is needed in summer but in winter it
serves only as a preheat for our natural gas water heater.
12.5 Active Solar Space Heat

Solar home heating is usually achieved with passive solar systems that are less
expensive (Section. 12.6). However, it is instructive to consider the more elegant
active solar systems that pump solar heated water to store in hot water tanks for
distribution to radiators. Active space heating systems are less competitive than
active hot water systems (Section. 12.4) because the home heating investment is
returned only during winter, while hot water for showers is used the entire year.
And, active hot water systems take advantage of the stronger summer sun, while
active space heating uses only the reduced winter sun. Less sophisticated, passive
solar heating systems, which rely only on glass and mass, are viable since houses
are heated only to a maximum of 70◦ F, as compared to 100–120◦ F hot water. We
assume the following for active solar home heating:
r thermally tight house with loss rate dQ/dt = (640 Btu/h ◦ F)T
r cold California day of 50◦ F
r η = 50% of daily integrated flux I = 2000 Btu/ft2 day.
P1: GFZ
Daily heat loss is
Q/day = (640 Btu/h ◦ F)(24 h/day)(15 ◦ F day) = 2.3 × 105 Btu/day. (12.30)
This gives a collector size
A = Q/ηI = (2.3 × 105 )/(1000) = 230 ft2 , (12.31)
made up of a eleven 3-ft × 7-ft panels. The total system costs some $15,000 unless the
owner assembles and installs the system. A properly designed passive solar home
of glass and mass can be built for much less money and without the maintenance
problems of hot water systems.
Active hot air systems use rocks to store heat. Rocks that are properly sized
have a thermal time constant to releasing most of the heat in about 5 evening
hours. Rocks have only 20% the specific heat per unit weight (c = 0.2 Btu/lb ◦ F)
of water because rock molar weight is greater than that of water. The highest vol-
ume heat-density storage is obtained with Glauber salt (Na2 SO4 · 10H2 0), which
has a liquid-solid phase transition at 90◦ F. Glauber salt releases 104 Btu/lb when
it freezes, as compared to 80 Btu/lb for water.2 Because of the phase transition,
Glauber salt needs only 20% of water’s storage volume. This allows its use in thin
Venetian blinds or other surfaces that are accessible to sunlight. If a salt with a 75◦ F
phase transition ever becomes viable, it will be warmly received by the energy
community.
12.6 Passive Solar Thermal Flywheel

12.6.1 Window Gains and Losses
Glass collects solar energy and traps infrared heat for passive solar’s glass plus mass
advantage. The energy gains and losses through south-facing windows in winter
are now calculated to evaluate the efficiency of passive solar energy. We begin again
with a favorable location in a moderate climate, San Luis Obispo, California, which
has few cloudy days in winter. We assume the following:
r double-glaze window, U = 0.5 Btu/ft2 h
r 45◦ F outside temperature in winter
r 90% transmission through
r winter flux, south-facing window, s = (270 Btu/ft2 h) sin(2π t/T) and T/2 = 10 h.
The daily heat loss is
Qloss /A = UTt = (0.5 Btu/ft2 h)(65◦ F − 45◦ F)(24 h) = 240 Btu/ft2 day. (12.32)
2
Liquid Glauber salt has specific heat c l = 0.68 Btu/lb ◦ F and its solid phase has c s = 0.46
Btu/lb ◦ F. Glauber salt has a density of 91.3 lb/ft3 , 1.5 times heavier than water.
P1: GFZ
12.6. Passive Solar Thermal Flywheel 311
The daily solar energy gain is
Qgain /A = 0.9sv T/π = (0.9)(270 Btu/ft2 h)(20 h)/π = 1550 Btu/ft2 day. (12.33)
The ratio Qgain /Qloss = 1550/240 = 6 is favorable and could be further improved
by using drapes or R11 Venetian blinds at night, or by R4 windows. Passive solar
heating can be used in more severe climates as a partial energy source that reduces
energy demand.
12.6.2 Thermal Flywheel

Adobe buildings are similar to engine flywheels, which smooth vibrations from in-
ternal combustion explosions. Massive adobe construction smoothes—or flattens—
the temperature cycle of the Southwest by bringing excess day heat for cool nights,
and excess night coolth to moderate warm days. Prior to the oil embargo, light,
2-by-4 construction without mass or insulation failed to smooth the temperature
cycle.
To illustrate this idea, the thermal time constant of 9-in diameter water pipes
is determined to see if these pipes can adequately shift day heat into the night.
To simplify the mathematics we idealize outside temperature as a square wave
cycle with outside temperatures that jump 20◦ F above 70◦ F room temperature for
12-h days, then 20◦ F below room temperature for 12-h nights (Fig. 12.9). How long
will the water tubes retain warmth? Will the tubes slowly drop to a reasonable
temperature of 70◦ F over a 12-h period? Will the time constant τ be too short as
seen by curve one (house with little thermal mass and little insulation) or will it
give pleasant curve two (adobe or Trombe-wall house)?
We ignore small temperature variations over storage volume, giving a stored
energy
Q = WcT, (12.34)
Figure 12.9. Thermal flywheel house. Curve (1) represents a house with little thermal mass
and Curve (2) represents an adobe or Trombe-wall house.
P1: GFZ
where W is water weight, c is water specific heat, and T is temperature difference

between water and room. For simplicity, the temperature differential T = Ttube −
Troom is denoted below as T, the temperature above room temperature. Heat loss
from stored energy is the time derivative of Q, that is
Q̇ = Wc Ṫ. (12.35)
Heat loss comes mostly from radiation and convection:
Q̇ ∼
= A(Uconv + Urad )T = AU total T, (12.36)
where A is tube surface area for a one-foot length and Utotal is the sum of radiation
and convection U-factors. Equating the surface loss rate to the loss rate of stored
energy inside gives
Q̇ = AU total T = −Wc Ṫ, (12.37)
which has an exponential solution, T = To e −t/τ with a thermal relaxation time
τ = Wc/AU total . (12.38)
Large heat capacity (Wc) gives a long relaxation time, while a large loss conduc-
tance (AUtotal ) gives a short relaxation time. The numerator weight × specific heat
(Wc) is proportional to stored energy, and the denominator area × transmittance
(AUtotal ) is proportional to energy loss rate of stored energy. The value of τ for 9-in
water tubes is
τ = Wc/AU total = (28 lb)(1 Btu/lb ◦ F)/(2.4 ft2 )(1.0 Btu/ft2 h ◦ F) = 12 h, (12.39)
where weight is 28 pounds/ft, area is 2.4 ft /ft with Utotal = 1. The 12-h relaxation
2
time is the correct choice to effect transfer of day heat to the night. An alternative
geometry entails flat panels (problem 12.16).
The energy stored per foot of tube, heated to 80◦ F, is
Q/ft = WcT = (28 lb/ft)(1 Btu/lb ◦ F)(80◦ F − 70◦ F) = 280 Btu/ft. (12.40)
This energy/foot is shed overnight,
Q/ft = AU total Tt = (2.4 ft2 )(1 Btu/ft2 h ◦ F)(80◦ F − 70◦ F)(10 h) = 240 Btu/ft.
(12.41)
The heat loss through the house envelope during a 12-h night is
Q/night = (600 Btu/h ◦ F)Tt = (600 Btu/h ◦ F)(70◦ F − 50◦ F)(12 h)
= 1.4 × 105 Btu. (12.42)
The length of tube needed to replace the lost heat is
(1.4 × 105 Btu/night)/(240 Btu/ft night) = 500 ft. (12.43)
A 1500-ft2 house with eight rooms requires 60 ft of water tubes per room. By
“insulating before you insolate,” tube length can be reduced. The solar energy
P1: GFZ
Problems 313
Figure 12.10. Denver house in the winter with R21 walls and 24-in adobe. Captured heat
is retained with only slight heating when temperature falls outside of the comfort zone
(Haggard et al., 2000).
available to heat the water is about 50% of the integrated solar flux 2000 Btu/ft2 day:
Qsolar /ft = ηI A = (0.5)(2000 Btu/ft2 day)(0.75 ft2 /ft) = 750 Btu/ft, (12.44)
which is more than sufficient to raise the tube temperature to 80◦ F with 280 Btu/ft.
Figure 12.10 shows simulation data for 24-in adobe walls with R21 insulation in
Denver on a winter day.
Problems
12.1 Dutch sun. (a) What is the solar angle θ at noon in Groningen, the
Netherlands, at 52◦ N latitude on December 21? March 20? June 21? (b) How
many air masses do sunrays pass through at noon and 3 PM on these days?
(c) What are horizontal and vertical solar fluxes for these days at noon and
3 PM using a sine curve?
12.2 Curve fitting. Plot a sine function for θ at noon for solstice/equinox days
of problem 12.1 to determine θ at solar noon on July 23, August 23 and
November 23?
12.3 Warmed water. (a) What solar flux is incident on a horizontal surface on June
21 at noon in Seattle (48◦ N) and Miami (26◦ N)? (b) Sun rays are absorbed in
P1: GFZ
black-ink water baths, 0.5-ft deep. What is temperature rise after 1 h, ignoring
heat losses, and allowing convection and radiation losses?
12.4 Integrated solar flux. What is integrated daily flux in the water tanks in
Seattle and Miami (problem 12.3) on June 23, when Miami has a 13-h day
and Seattle has a 16-h day? (b) How much does water temperature rise,
ignoring heat losses?
12.5 Elevation correction. (a) Air density decreases as e −h/H where h is elevation
and H is atmospheric height of 7 km. Correct the equation for s1 solar flux
to take into account site elevation. (b) Both Denver and Philadelphia are at
40◦ N latitude, but Denver is a mile high while Philly is at sea level. What is s1
on June 23 at noon for these two cities?
12.6 Gain/loss of south windows. What is energy gain/ft2 through south-facing
windows in Chicago on December 22 with 10 h of sunlight? Assume 20%
reduction in solar flux from clouds. How does this gain compare to heat loss
if outside temperature is 20◦ F? Assume double-glaze windows with U = 0.5
and 0.9 transmission.
12.7 Chicago blinds. Redo problem 12.6 but use R11 blinds during 14 h of dark-
ness.
12.8 Atascadero passive solar house. This house (35◦ N) proved that passive solar
can heat a house in winter without fuel. Water is placed in a covered glass
tank on the roof of a one-story house. The bottom metallic surface of the tank
acts as the ceiling of the house. This allows 10-h sunny days to heat the water.
In the late afternoon a motor pulls a thermal resistant cover over the top of
the tank to prevent night losses. Determine water depth for a temperature
rise of 10–20◦ F with 50% solar efficiency. The house has 1000 ft2 of roof with
lossiness of 300–500 Btu/h ◦ F and evening temperatures of 40–50◦ F over 14-h
nights.
12.9 Atascadero house in summer. Can the Atascadero house (problem 12.8)
comfortably survive 90◦ F, 14-h summer days? The motor covers the tank by
day and uncovers it by night when it is cooled by 55◦ F night air.
12.10 Night spray cooling. Night spraying of water on a low slope roof cools
the water by evaporation and radiation to 5–10◦ F below minimum night
air temperature. If 50% of the cooling can be captured, how much spraying
would it take to cool the Atascadero house, problems 12.9 and 12.10.
12.11 Phase transition storage. What is stored energy per lb and per ft3 for Glauber
salt (Section. 12.5) and water. Assume Glauber salt and water are heated from
70◦ F to 91◦ F.
12.12 Thermal relaxation. It takes 10 h for a 300-pound water drum to drop 63%
from its heated temperature to room temperature. What are the relaxation
times of 100-pound and 1000-pound barrels of similar shape?
12.13 Solar variations. (a) Determine the vertical integrated daily flux in Seattle
(47◦ N) on June 21, September 23/March 20, and December 21, while ignoring
clouds. What is θ on these three days at noon? (b) Compare your results on
annual basis with Fig. 12.7 for direct plus scattered flux. What is the scattered
flux, assuming our calculations and the table are correct?
P1: GFZ
Bibliography 315
12.14 Solar hot water. An 80-gal tank is heated on a clear spring day with a max-
imum solar flux of 270 Btu/h ft2 over a 12-h day with 40 ft2 of collectors at
50% efficiency. The 5-foot high tank has a radius of 1.5 ft and is surrounded
by R-6 insulation. Assuming no hot water is used during the day, what is
the temperature of the tank after 24 h if it was 60◦ F at 7 a.m.? The average
outside temperature during the day was 60◦ F.
12.15 South-facing glass. What area of south-facing glass is needed to heat houses
at 35◦ N latitude on December 21 with loss rates of 500T and 2000T Btu/h?
It is 45◦ F outside and the balance point of the building is 65◦ F. Assume 50%
passive solar efficiency.
12.16 Roof overhang. A house at 35◦ latitude needs an overhang to block light
in summer and transmit in winter. Design an overhang to cover a 3-m high
window with a 1-m wall above it.
12.17 Solar total. How does human energy use of 400 quads/year (2001) compare
with the amount of solar energy absorbed by Earth with an albedo (reflec-
tivity) of 0.3?
12.18 Flat panel passive solar. Use Eqs. 11.39–11.44 to analyze two-sided, flat water
panel collectors of 6–12 in thickness.
Bibliography
Adams, E. (Ed.) (2000). Alternate Construction, Wiley, New York.
Anderson, B. (1990). Solar Buildings and Architecture, MIT Press, Cambridge, MA.
Carriere, D. (1980). Solar Homes for a Cold Climate, Scribners, New York.
Haggard, K., P. Cooper and J. Rennick (2000). Alternate Construction, E. Adams (Ed.), Wiley,
New York.
Hsieh, J. (1986). Solar Energy Engineering, Prentice Hall, Englewood Cliffs, NJ.
Johnson, T. (1981). Solar Architecture: The Direct Gain Approach, McGraw-Hill, New York.
Kreider, J. (Ed.) (1981). Solar Energy Handbook, McGraw-Hill, New York.
Meinel A., and M. Meinel (1977). Applied Solar Energy, Addison Wesley, Reading, MA.
Montgomery, R. (1977). Solar Decision Book: A Guide to Heating Your Home by Solar Energy,
Wiley, New York.
Office Technology Assessment (1978). Application of Solar Technology to Today’s Energy Needs,
Sayigh, A. and J. McVeigh (1992). Solar Air Conditioning and Refrigeration, Oxford Univ. Press,
Oxford, UK.
Stine, W. (1985). Solar Energy Fundamentals and Design, Wiley, New York.
Swisher, J. (1985). Measured performance of passive solar buildings, Ann. Rev. Energy
Environ. 10, 201–216.
P1: GFZ
13
Renewable Energy
“The rational herdsman concludes that the only sensible course for him to purse
is to add another animal to the herd. And another, and another. . . Therein is the
tragedy. Each man is locked into a system that [causes] him to increase his herd
without limit—in a world that is limited . . . . Freedom in a commons brings ruin
to all.”
G. Hardin, Science 162, 1243, 1968
13.1 Sustainable Energy

Sustainable energy sources are nondepletable and environmentally benign. A re-
newable energy source is nondepletable by definition. A rigorous definition of
“sustainability energy” cannot be given, however, because each option has some
drawbacks that may not be apparent at first blush. For example, fossil fuels are not
sustainable for large modern economies and their use impacts the environment.
Fossil fuels probably would almost be a sustainable resource if world population
were only 1 million. Ultimately a definition of “sustainable energy” has to consider
the carrying capacity of Earth (Section. 10.5).
Garrett Hardin’s 1968 essay, “Tragedy of the Commons,” tells a story about en-
vironmental impacts to a shared, common green. When a herdsman adds an addi-
tional cow, the impact of its grazing is not immediately noticed since it is spread
among all the commons:
The counter argument is that technology allows nations to become wealthy and
mitigate damage better than poorer nations. Considerable mitigation is possible
if cheap electricity and renewable sources are competitive. This chapter considers
renewable energy from sun, wind, biomass, wave, geothermal, tidal and more.
Usually not considered are fission, which could be renewable if uranium extraction
from the oceans were to become cheap, and fusion, which is not yet economical.
Additional sources of energy from enhanced end-use efficiency technologies are
discussed in Chapter 14. From 1973 to 1995, enhanced end-use efficiency in the
United States saved more than 20 quads/year, far outpacing the additional 1–2
quads/year of additional renewable energy. However, after 25 years of research, it
now appears that commercialization of some renewable energy is ready for growth,
316
P1: GFZ
13.2. Photovoltaic Solar Power 317
Table 13.1. US Renewable Energy 2004

hydro geo wind solar bio ReE total E total
1973 quad/year 2.9 0.04 0 0 2.3 5.2 76.6
2004 quad/year 2.7 0.3 0.14 0.06 2.9 6.0 99.7
2004 avg. GWe 31 1.6 1.6 0.1 6.8 41 450
Renewable energy (ReE) production, total energy (E) use (quads/yr) and average electricity use (GWe).
(See Ch. 10, data from EIA, 2006)
lead by the wider adoption of wind energy. The US Senate passed an amendment
in 2002 that mandates 10% of electricity be produced by investor-owned utilities
from renewable energy sources.
Hydroelectric power is the largest US renewable energy source that has remained
relatively constant over the last two decades at an average of 35 GWe , but a dry
year can cut this amount by 30%. US wind energy capacity is projected to grow
at 6%/year (EIA, 2003), building on the 2004 average power of 1.4 GWe (0.2% of
electricity). Wind energy has experienced 30%/year growth recently in Europe,
particularly in Germany and Denmark. Wind’s cost is almost competitive with
the more flexible combined-cycle gas turbines driven with natural gas, but energy
storage is an issue if the fraction of wind energy becomes significant. Geothermal
(not strictly renewable) has grown to an average power of 1.6 GWe in the United
States, but its future growth depends on developing “hot-dry rock” sources. Look-
ing over the past, one finds there have been large over estimates in the early 1970s
on the growth of nuclear power (Atomic Energy Commission (AEC), 1200 GWe
by 2000) and renewables (Interior Department, 395 GWe of geothermal by 2000).
The dropping costs of renewables auger hope that they will play a large role in the
future. See Table 13.1 for US renewable energy data.
13.2 Photovoltaic Solar Power

Globally, photovoltaic (PV) cells produce about 300 MWpeak during peak sunshine,
with the United States producing about 40 MWpeak . PV capacity is projected (EIA)
to grow 20%/year. Whereas traditional fossil and nuclear plants are used 60–90%
of the time (their capacity factor), solar plants have a capacity factor of only 20%
due to the solar cycle. The solar capacity factor is determined from a PV’s rated
power at a peak solar flux of 1 kW/m2 , as compared to the US average (24-h /day)
solar flux of savg = 0.18 kW/m2 . This ratio gives the PV capacity factor of
CF PV = savg /speak = 0.18 kW/m2 /1.0 kW/m2 = 20%. (13.1)
High altitudes and clear skies raise CF in places like Albuquerque, which receives
70% more sun than Omaha. The land area needed to develop US average power of
400 GWe with 15% efficient photovoltaics is
A = PUS /savg η = (4 × 108 kW)/(0.18 kW/m2 )(0.15) = 1.5 × 1010 m2 = 15, 000 km2 .
(13.2)
P1: GFZ
318 13. Renewable Energy
One way to obtain this area would be to dedicate large regions to PV plants. To
save on dedicated land, PVs could be installed as roof tiles on buildings, giving
decentralized electricity that is independent of utilities. The main impediments for
a transition to PV electricity are high capital costs and the lack of inexpensive elec-
trical storage for nighttime use. However, PVs can reduce daytime peak electrical
power demand, which is 2–3 times night demand, without the need for storage
(unless it is cloudy day). Such an approach matches a solar supply cycle with a
peak-power demand cycle. Solar electricity could also be used to make hydrogen
for cars (Section 15.5). Solar electricity could pump water by day into elevated
reservoirs that become nighttime hydro plants. In small villages away from electri-
cal grids it usually is cheaper to using PVs to pump water, as compared to on-site
diesel generators. One way to lower PV costs is to use concentrators, which increase
solar flux by a factor of 10–20 to obtain peak solar flux of 10–20 kW/m2 . This re-
duces PV cost per peak watt, but adds the cost of the concentrator and tracking
mechanism (Section 12.2).
13.2.1 PV Efficiency
We estimate the efficiency of a silicon PV cell using these assumptions:
r silicon bandgap E gap = 1.1 eV
r open circuit voltage of 0.6 V, indicating that 45% of energy (0.5 V/1.1 V) goes to
waste heat
r solar spectrum with λ 0.4–0.8 μ (1.6–3.1 eV)
r 90% PV cell area is useable
r light is 10% reflected and 90% absorbed.
The efficiency of a PV cell depends on the relationship between the energy dis-
tribution of photons from the sun and the semiconductor’s bandgap. Photons with
energy less than E gap are not absorbed, since they cannot promote electrons from
the valence band to the conduction band. Photons with energy greater than E gap
are absorbed, but the excess energy beyond the bandgap energy appears as heat
from transitions from excited states to the conduction band. The efficiency of a PV
is maximized if the bandgap is chosen to accommodate the solar spectrum. A PV
with a small E gap mostly produces heat, while a PV with a large E gap absorbs few
photons.
For the case of silicon, each absorbed photon contributes bandgap energy of
1.1 eV (λ = 1.1 μ), making the entire visible spectrum available with the average
solar photon carrying 2.3 eV. The maximum efficiency of silicon PVs is
η = (0.9 area)(0.9 absorption)(0.55 non-loss)(1.1 eV Si)/(2.3 eV solar) = 21%.
(13.3)
This is consistent with the best laboratory value of 24.7%: Commercial modules
obtain η = 12−15%. It is generally believed that thin film or amorphous silicon
systems will be needed to make PVs economically competitive with natural gas.
P1: GFZ
13.2. Photovoltaic Solar Power 319
It is important to have long-lasting encapsulant for PVs. The 6-MW Arco PV plant
in San Luis Obispo had an early retirement because the ethyl-vinyl-acetate be-
came partially cloudy. Thin films have obtained η = 10−19% in the lab and up to
10% when placed in commercial modules. Amorphous PVs have obtained 6–9%
in modules, but aging of amorphous silicon may be a problem, as some of these
modules had a 20% drop in efficiency in 10 years. Gallium arsenide has a larger
band gap (E gap = 1.42 eV), closer to the average solar value, giving η = 32% in the
laboratory. Triple layer PVs, consisting of GaAs, GaInP2 (1.83 eV), and Ge (0.77 eV),
have obtained η = 34% (42% theoretical limit), but their high cost impedes com-
mercialization. Lastly, some new developments: Dyes in a conducting media can
collect light to transfer electrons to a conducting medium. Thermal resistant dyes in
a network of TiO particles gave 6% efficiency at 1 kW/cm2 , with only 4% reduction
in power over 1000 h. Developers claim they will be able to produce these PVs at
20% the cost of traditional silicon PVs. At low light levels a mixture of polymers
and CdSe nanorods gave η = 7% at 1 W/cm2 .
13.2.2 Cost/kWh from PVs

We estimate electricity costs in c/kWh at the busbar of a plant, using capital costs
in $/Wpeak , which is the electrical power when the sun is in the zenith at speak =
1 kW/m2 . We determine these costs at summer and winter solstices by assuming
the following:
r Silicon single crystal PVs cost of $5/Wpeak . In practice complete home units can
double this, but wide commercialization reduces costs to the $5 level/Wpeak after
rebates. It is hoped the costs of amorphous Si will be $1/Wpeak and thin-film
silicon will be $2/Wpeak . We ignore the 5% loss for conversion to 60 Hz.
r Clear skies and 25% clouds.
r PV collectors are seasonally adjusted to maximize sunrays at solar noon. This
takes full advantage of 10 h of sunshine at winter solstice and 14 h at summer
solstice at 30◦ N.
r Capital recovery rate of l0%/year (in constant dollars) to cover all costs except
energy storage and land (Section 16.1).
Capital cost, system lifetime, and capital recovery rate are relevant for PV system
marketing. The daily cost of 1 kWpeak at $5/Wpeak is
($5000/kWpeak )(0.10/year)(1 year/365 days) = $1.37/kWpeak -day. (13.4)
The area for 1 kWpeak at solar noon with η = 15% is
A = 1 kWpeak /ηs noon = (1 kWpeak )/(0.15)(1 kWpeak /m2 ) = 6.7 m2 . (13.5)
Since the cosine of the sun angle to the collector at noon varies slightly over a
month’s period, the cosine correction is ignored at solar noon. The solar flux perpen-
dicular to the aligned collector at solar noon on winter solstice at 30◦ N latitude with
P1: GFZ
solar angle θ = 30◦ + 23◦ = 53◦ using Eq. 12.16 is

snoon = so e −1/3cos53 = (1.367 kW/m2 )(0.58) = 0.79 kW/m2 . (13.6)
A similar calculation gives the flux at solar noon on summer solstice (θ = 30◦ −
23 = 7◦ ) of 0.98 kW/m2 . Under the assumption of sunny skies, the number of kWh
◦
generated per winter solstice day of 10 h sun (T = 20 h, Eq. 12.23) is

NkWh = snoon TAη/π = (0.79 kW/m2 )(20 h)(6.7 m2 )(0.15)/π = 5.1 kWh (13.7)
and NkWh = 8.8 kWh for a summer solstice day. For the case of $5/Wpeak costs, the
electricity costs $1.37/4.6 kWh = 27 c/kWh in winter. At summer solstice it costs
$1.37/8.8 kWh = 16 c/kWh, for an average of 22 c/kWh. There are also distribution
and other costs, which raise residential electricity by 4 c/kWh. For the case of 25%
clouds, the yearly average rises to 29 c/kWh at the busbar. If thin-film silicon is
realized at $2/Wpeak , the average busbar cost would be reduced to 9 c/kWh for
clear skies and 12 c/kWh for 25% cloudy skies. Amorphous silicon at $1/Wpeak
would drop the average to 4 to 6 c/kWh on a large system.
13.2.3 Solar Power Satellites

Solar flux above the atmosphere is eight times larger than the 24-h average flux
of 180 W/m2 . This advantage encouraged the United States, Japan and France
to design space stations to make solar electricity in geosynchronous orbit (GEO)
and beam microwaves to receivers on Earth. The receiver area needed in space to
obtain US average power of about 400 GWe is calculated with the assumption of
10% efficient electrical conversion, 70% efficient microwave generators and 75%
efficient receivers:
Aspace = PUS /so ηsolar ηmicrowave ηreceiver (13.8)
Aspace = 4 × 1011 W/(1367 W/m2 )(0.1)(0.7)(0.75) = 6000 km2 .

This represents a daunting receiver size of 50 miles by 50 miles. Excess Cold War
missiles have encouraged a new approach using low-Earth orbit (LEO) satellites
at an altitude of 200 km, increasing microwave-power density. LEO orbits reduce
microwave receiver area by a factor of 30,000 from the ratio of GEO to LEO distance
squared (36,000 km/200 km)2 . On the other hand, LEO satellites are on the other
side of the Earth half the time, therefore they must sell power to the whole globe.
13.3 Solar Thermal Power

13.3.1 Parabolic Trough
Solar thermal energy has been categorized as a “sleeping giant.” This section de-
scribes the conversion of solar flux for the purpose of driving turbines to make elec-
tricity. One approach to this process is used by Southern California Edison’s Luz
P1: GFZ
13.4. Hydropower 321
facility, which uses 1.5 million mirrors to produce 350 MWe . Compound parabolic
trough collectors increase solar flux by a factor of 10 on linear pipes, heating a
synthetic oil to 400◦ C. The problem of high pressures from high temperature steam
is greatly reduced by using synthetic oils as the primary fluid, which then passes
through a heat exchanger to make steam for turbines. Collector orientation is not
seasonally changed to take account of the sun’s 47◦ variation over the year. The
facility reports a cost of 12–17 c/kWh, with a thermal efficiency of η = 14%. We
will try to make sense of these numbers. The system cost $250/m2 ($3/We ) for
daytime average power, or
($250/m2 )/($3/We ) = 80 We /m2 . (13.9)
Using a daytime average solar flux of 400 W/m2 , the efficiency is
η = (80 We /m2 )/(400 Wavg /m2 ) = 20%. (13.10)
This is similar to the Luz results of η = 14%, which is 30% of Carnot efficiency for
Tcold = 320 K:
η = 0.3(1 – Tcold /Thot ) = 0.3(1 – 320K/670 K) = 16%. (13.11)
13.3.2 Power Tower

The power tower consists of many solar-tracking mirrors that project solar flux on
a boiler on a high tower. Solar flux is concentrated by a factor of 300 to 1500,
raising the boiler fluid to 500–1500◦ C. In 1999, Department of Energy Secretary Bill
Richardson reported that the 10 MWpeak power Solar Two power tower, using a
molten-salt working fluid (NaNO3 and KNO3 ), produced an average of 2 MWe
over a month’s period. One test of a system is the ratio of the average to peak
power. Thus, Solar Two’s ratio of 2 MWavg /10MWpeak = 0.2 was good, in fact, close
to the sun’s duty factor. This result implies excellent storage efficiency, which DOE
reported at 97%. In 2003, after three years of feasibility testing, Solar Two’s mirror
system was modified to monitor cosmic ray tracks in the sky.
13.3.3 Parabolic Dish

Tracking, parabolic reflectors have concentrated solar flux by a factor of 600–2000,
raising steam to 1500◦ C. Such a system has generated 10 kWe from a Stirling cycle
at 23% efficiency, which is 30% of Carnot efficiency.
13.4 Hydropower
Hydropower supplied 40% of US electricity in 1940, but now it supplies only 8%.
However, the hydroelectric contribution continues to be significant with an average
power of 30 GWe from 60 GWe capacity, along with 18 GWe pumped storage. Be-
cause of environmental impacts, it is not expected that the United States will build
P1: GFZ
more large hydroelectric plants. The largest US hydro plant, the Grande Coulee
Dam, Washington, develops 8 GWe with a dam 170-m high by 1.3-km wide. The
world’s largest hydro facility will be China’s Three Gorges Dam, which will dis-
place 1.1 million people with a 630-km reservoir when it is completed in 2009. Water
drops 181 m with an average flow of dV/dt = 1.1 × 104 m3 /s to develop power
with twenty-six 0.7 GWe generators:1
Phydro = ρghη(dV /dt)
= (103 kg/m3 )(9.8 m/s2 )(181 m)(0.9)(1.1 × 104 m3 /s) = 18 GWe . (13.12)
13.4.1 Kinetic Hydro

After the 1970s oil embargo there was a call to develop low-head hydro. These dams
have heights of less than 10 m. It was estimated that 49,000 US dams could develop
an average of 2 MWe , for a total of 110 GWe . This has happened only in a few cases
because environmental damage precludes it. However, British and Norwegian en-
gineers are developing underwater turbines with average power of 80–90 kWe as a
precursor to larger systems of 5 MWe . The turbines use three fiberglass-reinforced
blades to make electricity from ocean tidal currents. The maximum power that
could be developed is
PKE = 1/2(ρvA)v2 = (0.5)(1000 kg/m3 )(1.5 m/s)3 (314 m2 ) = 530 kW, (13.13)
where ρvA is water flow with velocity v = 1.5 m/s and blade area A = πr 2 =
π 102 m2 = 314 m2 . The 2002 British test gave 90 kWe for an efficiency of 90 KWe /530
kW = 17%.
13.4.2 Pumped Hydro Storage

Base-load plants operate 24 h/day, but they waste energy because only 30–50%
of daytime power is needed at night. The predominant method to save nighttime
base-load power is to pump water at night upwards to a reservoir on a high hill and
then release the water during the day to make hydroelectricity. By 2003, California
had built 4.2 GWe of pumped hydro storage as part of the US total of 18 GWe . The
Helms Project near Fresno pumps water during the evening to Pine Flat Reservoir,
485 m above Lake Wishon. During the day the water drops down a tunnel to
generate
P = ηρgh(dV /dt) = (0.9)(1000 kg/m3 )(9.8 m/s2 )(485 m)(265 m3 /s) = 1.1 GWe .
(13.14)
This is the power level needed to raise California Water Project water over the
San Gabriel Mountains into Los Angeles. The generators have an efficiency of 90%
to make electricity and 83% to pump water, for an all-over efficiency of 75%. The
1
(kg/m3 )(m/s2 )(m)(m3 /s) = (kg m/s2 )(m/s) = (newtons)(m/s) = F × velocity = watts.
P1: GFZ
13.4. Hydropower 323
system pumps 13 h at night and generates 11 h during daytime. The value of saved
energy used during daytime from Helms is
(1.1 × 106 kWe )(11 h/day)($0.12/kWh) = $1.4 million/day. (13.15)
The amount of water released from the Pine Flat Reservoir is
V = (265 m3 /s)(11 h)(3600 s/h) = 10.5 million m3 . (13.16)
Utilities in flat terrain can develop pumped hydro storage systems by using large
underground caverns instead of tall hills. A cavern 140 m on a side (2.7 million m3 )
located at a 2-km depth can develop 1 GWe for 10 h.
13.4.3 Helio Hydroelectricity

Heilio hydroelectricity generates power by channeling ocean water into below sea-
level depressions. “Helio” refers to the fact that in such a system output power
depends on the sun for evaporation. In the 1970s, a channel was proposed for
water to flow from El Alamein, on the Mediterranean Sea to the Qattara Depression
in northern Egypt. Peaceful nuclear explosions were considered for blasting the
80-km canal, but it was rejected in the end. More recently, a channel is envisioned
for water to flow from the Gulf of Aqaba to the Dead Sea.
Steady-state operation would be obtained when the flow rate of incoming water
is balanced with the evaporation rate of the new lake in the Qattara Depression.
Power is maximized when the lake is filled to a height that maximizes the product of
lake area and hydraulic head. Black dye is put in the water to enhance evaporation.
The Qattara project depends on the following parameters:
r maximized hydraulic head h× area A = 60 m × l2,000 km2

r evaporation rate dh/dt = 1.8 m/year
r efficiency of 80%.
The rate of evaporation from the Qattara lake would be
dV /dt = (dh/dt)A = (1.8 m/year)(1 year/3.2 × 107 s)(1.2 × 1010 m2 ) = 680 m3 /s,
(13.17)
which is 7% of Mississippi River flow. The helio hydroelectric power from the
depression would be
PHHE = (dV /dt)ρghη = (680 m3 /s)(103 kg/m3 )(9.8 m/s2 )(60 m)(0.8) = 320 MWe .
The deeper and smaller Dead Sea depression (h = 400 m, A = 750 km2 ) would
give PHHE = 130 MWe .
P1: GFZ
13.5 OTEC and Thermoclines

Oceans between the tropics of Cancer and Capricorn absorb considerable solar
energy, generating a 20◦ C difference between the surface (25◦ C) and 1-km depth
(5◦ C). This temperature difference was used in a shipboard demonstration to gener-
ate electricity. Once generated, the electricity can be sent by cable for use elsewhere,
or it can make hydrogen at sea. The Ocean Thermal Energy Conversion (OTEC)
process transfers warm surface water to a heat exchanger that vaporizes liquid
ammonia, the working fluid for this process. Ammonia vapor spins a turbine, pro-
duces electricity, and is condensed to liquid ammonia using the cold 5◦ C water
from below.
13.5.1 Boiler Depth

Heat transfer in the boiler heat exchanger and the condenser is most efficient when
the membrane thickness between water and ammonia is minimized. This can be
accomplished by choosing a depth that balances outward ammonia vapor pres-
sure with inward ocean pressure. The vapor pressure of NH3 is 7 atmospheres
(700 kPa) at 15◦ C. Each meter of ocean depth raises the pressure ρg(1 m) = (1000
kg/m3 )(9.8 m/s2 )(1 m) = 104 Pa/m. The depth D where water pressure equals
ammonia pressure is
D = (700 kPa)/(10 kPa/m) = 70 m.
13.5.2 Water Pumping Rate

The pumping rate for an OTEC plant is considerable. We assume the boiler heat
exchanger uses the 25◦ C warm water to heat the boiler fluid to 23◦ C. We also assume
the condenser uses 5◦ C cold water to cool the working fluid to 7◦ C. The amount of
heat delivered from 1 kg of hot water is
Qin /kg = mcT = (1 kg)(4.2 kJ/kg ◦ C)(2◦ C) = 8.4 kJ/kg. (13.18)
The OTEC efficiency is about 30% of Carnot efficiency:
ηOTEC = (0.3)(1 − Tcold /Thot ) = (0.3)(1 − 278/298) = 2%, (13.19)
where Tcold = 5◦ C (278 K) and Thot = 25◦ C (298 K). This is a low efficiency, but OTEC
fuel is free if we ignore other costs. The rate of work per kilogram of hot water is
W/kg = ηOTEC Qhot /kg = (0.02)(8.4 kJ/kg) = 170 J/kg. (13.20)
To obtain 1 GWe , it is necessary to circulate hot water through the boiler’s heat
exchanger at a rate
dm/dt = (109 J/s)/(170 J/kg) = 6 × 106 kg/s. (13.21)
For comparison, it is of interest to calculate the water pumping rate of a nuclear
power plant, which has a higher efficiency and a higher exhaust temperature. A
P1: GFZ
13.5. OTEC and Thermoclines 325
1-GWe nuclear plant rejects heat to the ocean by heating water from 12◦ C to 29◦ C. At
efficiency η = 1/3, coolant is needed to remove 2 × 109 W2 . This requires a cooling
water rate of
dm/dt = Pcold /cT = (2 × 109 J/s)/(4200 J/kg◦ C)(17◦ C) = 2.8 × 104 kg/s,
(13.22)
which is 0.5% of the OTEC rate. The OTEC demonstration was mounted on a ship
with its long tube, but it was not competitive, in spite of the strong support by
senators from Hawaii and Florida.
13.5.3 Fresh Water Thermoclines

Lakes in the Sierra Mountains have surface-to-bottom summer temperature differ-
ences as large as 19◦ C, sufficient to make electricity. Such temperature differences
have an effective pressure head h that is equivalent to the height of hydroelectric
plants. Equating the gravitational potential energy due to h to the stored heat energy
gives
mgh = mcT, (13.23)
where m is the mass of water, g is the acceleration of gravity, and c is the specific
heat of water (4200 J/kg ◦ C). Solving for effective head per degree of temperature
(h/T) gives
h/T = c/g = 427 m/◦ C = 780 ft/◦ F. (13.24)
◦
Thus, a 15 C difference is equivalent to a 6.4 km hydro head. However, ther-
moclines have temperature differences greater than 10◦ C for only 5 months. In
addition, thermoclines operate at low efficiencies, which we take into account by
reducing available thermal energy by a fraction f C times the Carnot efficiency,
f C T/T. The reduced equivalent hydro head h’ is obtained from
mgh = (mcT)( f C T/T), (13.25)

h = (c/g)( f C (T) /T) = (427 m/K) f C (T)2 /T.
2
(13.26)
Thus, the reduced equivalent head h’ is proportional to T 2 . The reduced h’
head for T = 15 K, T = 300 K, and f C = 0.3 (30% Carnot) is still considerable at
100 m:
h = (427 m/K)(0.3)(15 K)2 /300 K = 100 m. (13.27)
13.5.4 Salt Gradient Solar Ponds

Shallow ponds of very salty water have high-density gradients, which can have
temperature differences as large as 80◦ C, sufficient to produce electricity. In Israel,
2
If η = 1/3, then Ẇ = 1 GW = η Q̇in = (1/3) Q̇in . Thus, Q̇in = 3 GW. First law gives Q̇out =
Q̇in – Ẇ = 3 GW – 1 GW = 2 GW.
P1: GFZ
engineers took advantage of this phenomenon to build a salt gradient solar energy
system with a 1–4 m thick bottom layer of 20% salt content. The 1-m wide middle
zone has a density range of 1.0 g/cc at the top to 1.2 g/cc at the bottom. At the top of
the pond, a transparent zone is formed by rain. The top and middle zones transmit
solar radiation to heat the bottom zone, and the middle zone traps the upward
bound infrared from the bottom zone. The Israeli pond, which is located near the
Dead Sea, transferred 20% of incident solar energy to a Rankine-cycle turbine that
made electricity with an efficiency of 8.5%. The Dead Sea plant developed 200 kWe
with a combined solar-turbine efficiency of 1.5%. The United States has a similar
70-kWe facility near El Paso, Texas, since 1986. Membranes have been used to
prevent interlayer convection.
13.6 Wind Power

Wind power is the fastest growing renewable source of energy. It increased 25%
during 2000 when 4.5 GWe was added, raising the global total to 18.5 GWe (13.6
GWe in Europe, mostly in Germany and Denmark, and 2.6 GWe in the United States,
mostly in California, Texas, and the Midwest). During the next 2 years, US wind
capacity rose to 4.6 GWe (average power 0.7 GWe ). Taking advantage of steady,
offshore winds, Europe had built 86 MWe offshore by 2000, and it expects to have 2.4
GWe by 2006. Wind power is projected to grow at 6%/year in the United States due
to dropping costs and government tax credits of 1.8 c/kWh. Not everyone favors
wind power, as some communities have opposed construction of wind farms off
the coasts of Cape Cod and Long Island, mostly for aesthetic reasons.
Wind energy flux is its kinetic energy density (0.5ρv2 ) times the volume defined
by a 1-m2 area times the distance wind travels through the cross-sectional area in 1
s (wind velocity v). Thus, the flux of wind energy is proportional to wind velocity
cubed, v3 . A wind velocity of 6.5 m/s (23 km/h = 15 mph) has wind flux equal to
the 24-h US average solar flux of about 180 W/m2 :
Pwind /m2 = 0.5ρv3 = 0.5(1.3 kg/m3 )(6.5 m/s)3 = 180 W/m2 . (13.28)
There are locations in the Texas Panhandle and mountain passes, such as
Medicine Bow, Wyoming, where average wind velocities of 8–9 m/s generate wind
fluxes as high as 600 W/m2 . Many places in the Great Plains have power class-4
sites with average winds of 5.6–6.0 m/s at 10-m height. Since wind velocity
increases as the 1/7 power of height, class-4 velocities scale to 50-m height,
to give
ν50-m = ν10-m (50 m/10 m)1/7 = 1.26ν10-m = 1.26(5.6−6.0 m/s) = 7.0−7.6 m/s.
(13.29)
This raises class-4 wind flux from 200 to 250 W/m2 at 10 m to 400 to 500 W/m2 at
50 m. Even higher velocities are available with stratospheric 100-mph winds, but
at reduced air density, which give fluxes over 1000 W/m2 .
P1: GFZ
13.7. Tidal and Wave Power 327
Textbooks list US wind power potential at 1015 W (1 million GW), with 100 GW
at “excellent” sites. The 1 million GWe potential is model-dependent, but we can
obtain this result by assuming an average of 10 m of useable height over an area of
the lower 48 states (1500 km × 5000 km). This gives a wind volume of
Vwind = (1500 km)(5000 km)(0.01 km)(109 m3 /km3 ) = 7.5 × 1013 m3 . (13.30)
The theoretical maximum efficiency of windmills is 59% (problem 13.16), but
operating windmills operate at 25% efficiency. The lower-48 potential for electricity
from wind power at 5 m/s is
Pwind = 0.5ηρv3 Vwind = (0.5)(0.25)(1.3 kg/m3 )(5m/s)3 (7.5 × 1013 m3 ) ≈ 1015 We ,
(13.31)
which is 10,000 times larger than the US average electrical power of 4 × 1011 We .
Electrical power generated from a 60-m diameter, 25% efficient windmill at a
class-4 site (50-m tower, v = 7 m/s) is
Pwind = 0.5ηρv3 Awind = (0.5)(0.25)(1.3 kg/m3 )(7 m/s)3 (π )(30 m)2 = 1.5 MWe .
(13.32)
The v3 power dependence encourages the choice of sites with fast winds, but it
also magnifies the effects of velocity fluctuations. For example, the 7-m/s average
velocity can be represented by equal parts of 3 m/s, 7 m/s, and 11 m/s. The average
of v3 , or v3 , in this case is
v3 = (33 + 73 + 113 )/3 = 1701/3 ≈ 1.7(7 m/s)3 = 1.7vavg
3
. (13.33)
This calculation shows that velocity fluctuations enhance average power by a factor
of 1.7 when compared to power at the average velocity. In any event, turbines do
not operate when velocity is very low, as they are designed for higher velocities.
Because of this, the industry practice is to give wind machines a duty factor of about
25%. For wind farms it is important that one wind machine not shadow another
since this reduces power. If turbines are placed with a separation of five windmill
diameters, power reduction can be as large as 30%, but if separation is 10 diameters,
reduction is only 10%. Wind machines priced at $2000/kW with a rated capacity
factor of 0.5 and a 10% capital recovery factor are almost competitive with natural
gas at 3–4 c/kWh at the busbar, as calculated here:
(0.1/year)($2000/kW)/(0.5 × 8760 h/year) = 4.5 c/kWh. (13.34)
13.7 Tidal and Wave Power

13.7.1 Global Tidal Power
The global tidal power potential is determined from the moon’s tidal torque, a kind
of break shoe that slows Earth’s rotation. The present Earth angular frequency is
ωpresent = 2π/(24h)(3600 s/h) = 7.3 × 10−5 rad/s. (13.35)
P1: GFZ
Past angular frequency is determined by counting daily growth rings of ancient

corals. Geologists determined that Earth had a rotation rate of 395 days/year during
the Devonian era, 350 million years ago, giving an angular frequency at that time
of
ωDevonian = (7.3 × 10−5 )(395 days/365 days) = 7.9 × 10−5 rad/s. (13.36)
This gives Earth’s angular deceleration,
α = (ωpresent −ωDevonian )/t = −(0.6×10−5 rad/s)/

(3.5×108 year)(3.2×107 s/year) = −5.4 × 10−22 rad/s2 . (13.37)
The tidal torque for this deceleration is given by
τ = I α = 0.4mEarth R2E α, (13.38)
where I is Earth’s moment of inertia (0.4mEarth R2E ). Under the assumption of a

constant-density Earth, the global torque is
τ = I α = 0.4(6.0×1024 kg)(6.4×106 m)2 (−5.4×10−22 /s2 ) = −5.3×1016 N m.

(13.39)
Tidal torque reduces Earth’s rotational energy at a rate
Ptidal = τ ωnow = (5.3 × 1016 )(7.3 × 10−6 ) = 4 × 1012 W = 4000 GW, (13.40)
which is close to the 2600-GW textbook value. The best US tidal resources are in
Maine, near the Canadian Bay of Fundy, which has a 20-MWe tidal plant. However,
the building of tidal dams blocks estuaries that are vital for spawning fish. (See
Section 13.4 on undersea turbines run with tidal power.)
13.7.2 Saint Malo Power Plant

The world’s largest tidal power plant was built in 1966 on the La Rance Tidal
Basin, Saint Malo, France. It is driven with a 7-m tidal range between high and low
tides, filling and emptying the 23-km2 tidal basin. The power plant operates at 90%
efficiency over the T = 6-h and 10-min tidal cycle (from high to low tide).
The potential energy of the filled tidal reservoir is
E potential = AH 2 ρg/2 = (23 × 106 m2 )(7m)2 (1000 kg/m3 )(9.8 m/s2 )/2 = 5.5 × 1012 J.
(13.41)
This energy develops power over the high to low tidal cycle period T (and the
low-to-high cycle):
Ptidal = ηE potential /T = (0.9)(5.5 × 1012 J)/(6.2 × 3600 s) = 0.22 GWe . (13.42)

P1: GFZ
13.7. Tidal and Wave Power 329
13.7.3 Wave Power

The first commercial wave power plant was destroyed by big waves off the coast of
Scotland in 1977, but efforts to harness wave power continue because the resource
is so abundant. The open ocean usually has calm, but choppy waves. But, as wind-
driven waves approach the shoreline, their velocity slows and their wave amplitude
grows and the waves become more complex.3 As the front half-cycle of the wave
further slows, the faster back half-cycle piles onto the front half-cycle. Waves break
when the wave height becomes greater than λ/7. It is the wave height that gives
the potential energy for wave power.
The wave power potential of Earth derives from the characteristics of its oceans:
Average waves have amplitude yo = 1 m, occurring every 10 s with frequency f
= 0.1 Hz, wavelength λ = 10 m, and velocity of 1 m/s. Waves beat against half
the world’s coastline of 4 × 105 km (the other half is protected by peninsulas and
islands). The potential energy E potential of waves contained in an area of beach width
z, and reaching toward the ocean one wavelength λ is
E potential /z = ρgλy2 = ρgλyo2 /2. (13.43)
The y2 factor is the average of the square of the wave height above y = 0; the first
y is proportional to potential energy and second y is proportional to the amount
of raised water. The second equation rewrites this, using y2 = yo2 /2 by averaging
over one-half a sine wave. The potential energy is about the same as kinetic energy,
doubling the possible energy available. This gives an approximate wave power per
unit length of beach ( f = 1/T= 0.1 Hz) of
Pwave /z = ρgλ f yo2 = (103 kg/m3 )(9.8 m/s2 )(10 m)(0.1 Hz)(1 m)2 = 10 kW/m.
(13.44)
Wave generators harvest only some of this, tapping either the kinetic or poten-
tial energy components. Global coastlines receive wave power at about the same
magnitude as global tidal power,
PEarthwave = Pwave L coast /2z = (10 kW/m)(4 × 105 km)/2 = 2000 GW. (13.45)
One device built by UK scientists has a sphere vertically attached to the top of
a spherical buoy. The two spheres are connected with a pipe to allow water to be
raised from the lower sphere to the upper sphere. The oscillating motion drives
water into the upper sphere to push against a piston, driving a generator. Wave
energy is captured with a flapper valve that allows water to flow up, but not down.
Other systems oscillate mass cabled from a buoy to spin a generator or to have
waves hit water turbines.
3
In shallow water with depth less than λ/2, wave velocity drops according to vwave = 3.1
m/s(d)1/2 , where depth d is in meters.
P1: GFZ
13.7.4 Compressed Air Machine

Japanese scientists built a wave machine that uses the rising motion of the wave
surface to compress air inside a chamber to push a bellows to drive a turbine. The
turbine responds to the work done on the gas essentially at constant4 pressure po =
105 Pa to yield
Vf
E =− po dV = − po (Vf − Vo ) = (105 Pa)(100 m3 − 150 m3 ) = 2.5 MJ.
Vo
(13.46)
The Japanese system parameters of Vo = 25 m2 × 6 m = 150 m3 and Vf = 25 m2 ×
5 m = 100 m3 accounts for an average 2-m rise of water. Wave power is Pavg =ηE/T
= 0.1(5 MJ)/10 s = 50 kW, during a 10-s period at 10% efficiency.
13.7.5 Dam-Atoll
Lockheed Corporation designed the Dam-Atoll to focus and convert kinetic energy
of waves into potential energy for turbines. The atoll is a floating shell with a
diameter of 80 m, concave side down. This geometry reduces the depth of waves
as they wash over the shell. The reduced depth slows the waves, raising the index
of refraction. The index of refraction is highest for waves closest to the center of
the atoll, making the waves spiral around the center, focusing them to the center.
Water-focusing is similar to the role of an optical lens, refracting waves toward the
region of slowest velocity. Focusing raises wave amplitudes to spin a turbine at the
center of Dam-Atoll.
Lockheed’s patent application claims an input wave power of 2.2 MW for waves
with an amplitude of 1 m and a period of 7 s, giving a projected average power
of 1 MWe . The reduced wave depth on Dam-Atoll’s surface redirects the kinetic
energy, raising wave amplitudes, giving 2200 kW/80 m = 25 kW/m of input power.
13.8 Geothermal Power

Radioactive decay and magna solidification that increase Earth’s inner core produce
an average geothermal heat flux of 0.06 W/m2 . The geothermal power potential at the
US surface is
Pgeo = (dP /dA)A = (0.06 W/m2 )(8 × 1012 m2 ) = 5 × 1011 W = 500 GW. (13.47)
Assuming no natural water source and an average gradient of 30 K/km, a 1-km
deep well with pumped water could heat homes geothermally. Similarly a dry well
several kilometers deep could drive a power plant with injected water from above.
Most of this power is not economically available. However, high-grade geothermal
4
Treating this more properly as an adiabatic compression gives a similar result.
P1: GFZ
13.8. Geothermal Power 331
sources near plate boundaries and recent volcanoes can be economically viable.
These more active locations have larger thermal gradients of 150–200 K/km that
produce steam above 150◦ C. The Geysers, 100-km north of San Francisco produce
50% of US average geothermal power of 1.6 GWe (2.8 GWe capacity). Modern
geothermal wells capture H2 S gas and reinject wastewater containing boron and
ammonia. The large resource in California’s Imperial Valley requires re-injection
of hot brine into the Earth for disposal. Along similar lines, engineers in Hawaii
have extracted heat from 650◦ C molten magma. It is estimated that the Hawaiian
source of 360◦ C steam is sufficient to produce 1 GWe for 50 years. Some geothermal
resources seem “sustainable” but others are depleted as heat removal surpasses
heat replenishment.
13.8.1 Wet Geothermal

Low-quality geothermal steam energy is converted to electricity by using large,
low-pressure turbines that extract heat in a two-step process. Initially steam and
water are separated at the first turbine; the steam drives a turbine and the hot
water is vaporized at reduced pressure to drive a second turbine. We estimate the
properties of a well assuming (1) 130 kg/s flow at 182◦ C (455 K) and 8 bars; (2)
available enthalpy, h = Uint + pV = 2.8 MJ/kg; and (3) condenser temperature of
47◦ C (320 K). The available thermal power is
Pgeo-t = (130 kg/s)(2.8 MJ/kg) = 365 MWt . (13.48)
At 50% Carnot efficiency, electrical power is
Pgeo-e = 0.5ηC Pgeo-t = 0.5(1 − 320 K/455 K)(365 MWt ) = 55 MWe . (13.49)
13.8.2 Dry Geothermal

Scientists at Los Alamos National Laboratory carried out feasibility studies regard-
ing exploitation of plentiful dry hotspots. Water was pumped into the earth for
geothermal heating and then pumped to the surface to cycle through a turbine.
The crust was hydraulically cracked at 2-km depths with high-pressure water to
increase surface area of the underground heat exchanger. Project data were as fol-
lows:
r entrance water of 8 kg/s at 35◦ C
r exit water of 8 kg/s at 170◦ C (dropped to 90◦ C over time) with 2% water loss
r pressure drop from 28 bars at entrance to 14 bars at exit
r 100-m distance between entrance and exit pipes.
The maximum thermal power obtained was
Pgeo-t = cT(dm/dt) = (4200 J/kg ◦ C)(170◦ C − 35◦ C)(8 kg/s) = 5 MWt .

(13.50)
P1: GFZ
At 30% of Carnot efficiency, the dry well could supply electrical power
Pgeo-e = 0.3ηC Pgeo-t = 0.3(1 − 308 K/443 K)(5 MWt ) = 0.5 MWe . (13.51)
13.9 Biomass Power

Biomass is a form of solar energy obtained from photosynthesis, as carbon dioxide
and water are transformed into glucose:
6CO2 + 6H2 O ⇒ C6 H12 O6 + 6O2 . (13.52)
Global photosynthetic processes yield 2.2 × 1011 tons/year dry biomass, 10 times
the global energy need. The efficiency for the conversion of solar energy into
biomass energy can be as high as 2% with certain kinds of kelp and 1.2% with
sugar cane, wheat, and soybeans. The potential to convert solar energy to biomass
energy is about
(0.01)(200 W/m2 ) = 2 W/m2 . (13.53)
Sugar cane has an average yield of 120 ton/hectare year, or 12 kg/m2 year, which
is partially sugar. In addition, the United States has 100 million cows, 60 million
pigs, 7.5 billion chickens and 290 million turkeys to transform plant biomass into
meat and waste biomass.
Biomass contributes 38% of total energy in lesser developed countries, but only
3% in industrialized states. Biomass contributed 2.9 quads, or 3% of US energy use
in 2001. On the other hand, it took 12% of the national energy budget to grow food
and wood. It takes about 6 J of primary energy to create 1 J of food product. Of US re-
newable energy sources, biomass contributes about the same as hydropower, but it
greatly surpasses wind and solar power. US biomass generated average of 6.8 GWe
in 2001, 75% from forest-product residues and 25% from municipal solid wastes.
13.9.1 Gasohol
Brazil found its economy almost bankrupt after the 1970s oil embargo because
of high-priced OPEC petroleum. However, Brazil has very favorable conditions to
develop a domestic alcohol industry, which let it reduce its dependence on gasoline.
Excellent growing climate, considerable land, and cheap labor were factors that
encouraged Brazil to create a massive industry, using 4% of its croplands to grow
sugar cane for ethanol. Sugar is converted into ethanol by fermentation, with little
loss of energy:
C6 H12 O6 ⇒ 2C2 H5 OH + 2CO2 . (13.54)
Since 1979, Brazil has not allowed pure gasoline-burning automobiles. About one-
half of Brazil’s cars run on pure ethanol and the other half use a 22% ethanol, 78%
gasoline mixture.
The US situation is not comparable since it has a shorter growing season, expen-
sive labor and less available land. Corn is used by the United States for its source of
P1: GFZ
13.10. Fusion Power 333
ethanol. However, corn does not contain sugar, but it does contain starch, which re-
quires further processing. There is also the question of using food resources to fuel
cars in a hungry world. The net energy balance appears to be neutral. According to
the Office of Technology Assessment (Office Technology Assessment, 1980), “The
sum of the heats of combustion of the fuels in cultivating, harvesting and drying
grain and converting it to ethanol is about the same as the heat of combustion of
the resultant ethanol (about 22 MJ/l).” More recent studies show a modest net gain
in energy from ethanol, but its drawbacks remain. However, a switch to ethanol
is a net gain in the sense that it reduces petroleum consumption, since energy to
make gasohol usually comes from coal or natural gas. Another option is methanol
(CH3 OH) from wood products, but obtaining it is more complicated than obtaining
ethanol because wood contains considerable cellulose, which is more difficult to
convert to a liquid fuel. We examine the subsidy to gasohol in Section 16.7.
13.9.2 Garbage Power

Garbage has an energy content of about 10 MJ/kg (4500 Btu/lb), about 40% the
energy content of oil and coal. An individual American produces 2 kg/day garbage
for a total of (2 kg/day)(365 day/year) = 0.7 ton/year. The United States produces
(280 M)(0.7 ton/year) = 200 Mton/year of municipal solid waste. This is about 10%
of the raw materials needed to run the United States at a rate of 2700 Mton/year.
The energy content in US garbage is
(200 Mton/year)(1010 J/ton) = 2 × 1018 J/year = 2 quad/year = 1 Mbbl/day,

(13.55)
which is 2% of the total energy budget. In spite of these favorable numbers, garbage
power has limitations. Much of the high-energy-content is in paper and plastic that
are now separated from garbage in most cities, lowering the quality of garbage
fuel. In addition, garbage must be presorted, air pollution control equipment must
contain noxious chemicals, and water content must be constrained. The nation
produces an average of 2 GWe from burning municipal solid wastes at 100 facilities.
Garbage and biowaste can be made into low-grade petroleum products, methanol,
and hydrogen. Indeed, petroleum was produced over the eons by slightly warming
biomass waste in the absence of oxygen. The same long-exploited process can be
used to pyrolyze waste biomass into tarry fuels. Surrounding biomass with heat
from geothermal wells or waste industrial heat, while denying oxygen, creates the
hydrocarbons.
13.10 Fusion Power

Physicists have long dreamed of using deuterium in the oceans to produce plentiful
and relatively clean “renewable” energy. Deuterium (0.015% of hydrogen) could
supply world energy needs for 50 billion years at the present consumption rate.
P1: GFZ
The hydrogen bomb reaction, which combines deuterium (D) and tritium (T) to
produce helium, might someday produce commercial electricity:
2
H +3 H ⇒ 4 He +1 n + 17.6 MeV. (13.56)
We first consider magnetic confinement fusion, which uses magnetic fields to con-
fine hot DT plasmas in a magnetic “bottle.” In 1997, the UK Joint European Torus
announced an output of 10 MW for 0.5 s from its tokomac machine, a result we will
analyze.5 We then consider laser-induced fusion, which uses lasers to implode and
heat DT pellets. Since this reaction is very quick, the energetic DT nuclei combine
before thermal velocities disperse them. This approach is termed inertial confinement
fusion (ICF) because the DT inertial mass allows for sufficient time for reactions to
take place. ICF drives hydrogen bombs, but it has also been touted as an energy
source.
13.10.1 Magnetic Confinement Fusion

When a deuterium beam passes through a region containing tritium, its particle
flux density is reduced because of DT reactions. For a thin slab of tritium, the
deuterium flux f D (particles/m2 s) of deuterons is reduced by
f D = − f D σ n T x (13.57)
where σ is DT cross-section, nT is tritium density, and x is slab thickness. This
integrates to give an exponentially decreasing flux,
f D = f Do e −nσ x . (13.58)
Deuterium flux is the product of deuteron particle density n D and deuteron
velocity, or f D = n D v. The DT fusion rate in a volume of slab area A and thickness
x is
RDT = f D nT σ (Ax) = nD nT σ v(Ax). (13.59)
The power density for DT reactions, each with energy E DT , is
PDT = E DT RDT /Ax = E DT nD nT σ v . (13.60)
DT plasma density n contains equal amounts of D and T, giving nD = n/2 and
nT = n/2. The distribution of plasma velocities contains high velocity ions that en-
hance DT cross-sections, which increase with temperature up to 0.1 MeV. At higher
temperatures, Bremsstrahlung radiation from accelerating electrons becomes ex-
cessive. The trade-off between a rising DT cross-section and Bremsstrahlung losses
dictates an operating temperature of some 50 million K (5 keV). At that temperature
the thermal average of σ v is σ v = 5 × 10−24 m3 /s. Combining these results gives
5
“Tokomac” is a Russian acronym for “toroid kamera magnit katushka” (toroidal chamber
and magnetic coil), invented by Andrei Sakharov. See Princeton Plasma Physics Lab site
(//ippex.pppl.gov) for simulations and data analysis.
P1: GFZ
a power density of
PDT = n2 σ vE DT /4, (13.61)
where n is plasma particle density. The 1997 UK result of 10 MW for 0.5 s probably
was observed at a region somewhat above the break-even point, when fusion en-
ergy gain equals Bremsstrahlung losses. Outside thermal energy can be supplied
with neutral beam heating and ohmic heating from external beams. The break-even
energy release (with 1/3 recovery heat) is defined as the Lawson criterion,
nτ = 6 × 1013 particles s/cm3 = 6 × 1019 particles s/m3 . (13.62)

where τ is confinement time in s. Using the Lawson criterion with the UK result of
10 MW for 0.5 s gives the minimum particle density n
n = 6 × 1013 particles s/cm3 /0.5 s = 1014 particles/cm3 = 1020 particles/m3 .

(13.63)
Since the US efforts obtained 5 × 1020 particles/m3 in the 1990s, we arbitrarily
increase n to 1021 /m3 . Using these values, the fusion power density is
PDT = n2 σ vE DT /4 (13.64)
PDT = (1021/m3 )2 (5 × 10−24 m3 /s)(17.6 MeV)(1.6 × 10−13 J/MeV)/4 = 3.5 MW/m3 .

This is consistent with the experimental result of 10 MW in the UK torus (3-m
outer radius, 1.3-m inner radius and 1-m height). The total energy produced is 5 MJ
(10 MW × 0.5 s), which is much larger than the 4 × 10−3 J, obtained from the inertial
confinement demonstration at Sandia’s Z-machine, described below.
13.10.2 Laser Fusion

The chair of the US government’s 1979 review panel on ICF, John Foster, concluded
that “We see no insurmountable roadblocks to the practical achievement of elec-
trical power generated by inertial confinement fusion.” In spite of this glowing
conclusion, ICF has always been used primarily to study nuclear weapon mate-
rials at high temperatures and pressures. As a CTBT signatory, the United States
is bound not to test nuclear weapons, but rather to rely on the Stockpile Steward-
ship Program to maintain the safety and reliability of its nuclear weapons. Toward
this end, DOE hopes to complete by 2008 the construction of the National Ignition
Facility (NIF), which uses lasers to heat and implode weapon materials like tri-
tium, deuterium, plutonium, and uranium. NIF will divide its 1.8-MJ beam energy
among 192 beams that will implode tiny spherical targets. NIF can also heat a cavity
to very high temperature, to radiate copious x-rays to implode small spherical tar-
gets. The tiny DT spheres are estimated to rise to 100 million K, hot enough to cause
fusion. In addition, uranium/plutonium compression data will give equations of
state information for compressing primaries and to examine aging plutonium. The
P1: GFZ
deuterium/tritium data will give equation of state information for boosted pri-
maries and hydrogen secondary stages. Some say that NIF should not be built
because it gives the perception that the United States is working toward new type
of nuclear weapons, while others say it is needed for stockpile stewardship and to
make weapons work more interesting.
Time to compress ICF. As stated already, the Lawson criterion for break-even energy
release from DT fusion is
nτ = 6 × 1013 particles s/cm3 . (13.65)
This suggests that a plasma density n = 6 × 1013 particles/cm3 must be confined
for at least one second so that the DT fusion energy can balance the energy to make
the plasma and overcome losses. The Lawson criterion serves as a rough guide to
predicting available time for ICF. Because the target spheres are very small, the
pressure of DT gas can be very large. As the vaporized glass surface of the pellet
is exploded, the radial reaction force implodes the remaining DT sphere. Since the
initial pellet is cooled to 18 K, we use the liquid density of hydrogen, nH-liquid = 4.5
× 1022 /particles s/cm3 . The minimum confinement time is about
tICF = (6 × 1013 s/cm3 )/(4.5 × 1022 s/cm3 ) = 1.3 ns, (13.66)
which is consistent with the pulse width of the 500-TW NIF,
t = E/P = (1.8 MJ)/(5 × 1014 W) = 3 ns. (13.67)
Lawrence Livermore National Laboratory states6 that NIF compression would
increase hydrogen density to “20 times that of lead,” If this were achieved, the
minimum confinement time would be correspondingly reduced. The fast ignitor
concept for ICF uses an initial short pulse of 0.1 ns to implode a DT capsule. When
the pellet is at a maximum density, a very intense (1019 –1021 W/cm2 ) second pulse
with a time width less than 0.01 ns ignites DT fusion. The second pulse creates hot
electrons (0.1–1 MeV) that raise ion temperatures to 5–20 keV, sufficient for fusion.
ICF pressures. Just prior to implosion, the molar density of liquid hydrogen is
nmole = n/NAvogadro = (4.5 × 1022 /cm3 )/(6.023 × 1023 ) = 7.5 × 104 mole/m3 .
(13.68)
The perfect gas law at 100 million K gives a DT pressure of
p = nmole Rgas T = (7.5 × 104 mole/m3 )(8.3 J/K mole)(108 K) ≈ 1014 Pa = 1 Gbar.
(13.69)
(In a related mechanical experiment, shock waves produced 2 × 1011 Pa by project-
ing a flat disk at 7 km/s on a cooled Pu target.). Some lasers have intensity (energy
flux) as high as 1021 W/cm2 , which gives a light pressure of
p = f /c = (1025 W/m2 )/(3 × 108 m/s) = 3 × 1016 Pa = 300 Gbar. (13.70)
6
LLNL Science and Technology Review, p. 5–11, July. 1999.
P1: GFZ
Pellet size. The size of a DT pellet is estimated from velocity of sound in the hot
pellet and minimum confinement time. The speed of sound in gas is
vsound = (constant)(T/M)1/2 , (13.71)
where T is temperature and M is molal mass. Since sound velocity in air at 0◦ C is
331 m/s, the average velocity of deuterons and tritons in hot plasma is
vs−DT = vs -273K [(TDT /TSTP )(Mair /MDT )]1/2
= 331 m/s [(108 /273)(29/2.5)]1/2 = 6.8 × 105 m/s. (13.72)
Since velocity of sound is similar to the diffusion velocity, the radius increases
during the confinement by
r = vs-DT tICF = (6.8 × 105 m/s)(1.3 ns) = 0.9 mm, (13.73)
which is similar to the 1-mm initial radius. The initial volume of the 1-mm radius
pellet is Vpellet = 4 × 10−3 cm3 . If 10% of DT is fused, the energy yield is
E = 0.1nVpellet E fusion
= 0.1(4.5 × 1022 /cm3 )(4×10−3 cm3 )(1.76×107 eV)(1.6×10−19 J/eV) = 5.1 × 107 J,
(13.74)
where E fusion = 17.6 MeV = 1.76 × 107 eV. The explosive yield of the pellet corre-
sponds to a mass of TNT of
ma ss = (5.1 × 107 J)(1 kton/4.2 × 1012 J)(9 × 105 kg/kton) = 10 kg, (13.75)
which is consistent with estimates for laser-pellet power plants. Particle beam ex-
citation might be more successful than laser fusion since particle beams are made
with 20–40% efficiency, as compared to the 5–10% efficiency to make a laser beam.
Furthermore, particle beams might couple more completely to the pellet.
Laser size. Complicated calculations are needed to estimate the laser size required
to achieve laser fusion. One would expect that the initial laser pulse ionizes DT to
create a sea of free electrons, much like an alkali metal. Since alkali metals reflect
light, the pulse timing and wavelength must be patterned to enhance absorption by
the pellet. Beyond complications of energy coupling, the effects from super-shocked
fusion hydrodynamics must be estimated. Nonetheless, we roughly estimate the
laser size needed to preheat DT for fusion. The number of moles of DT atoms in
the pellet of volume V = 4 × 10−3 cm3 is
Nmole = nmole V = (0.075 mole/cm3 )(4 × 10−3 cm3 ) = 3.0 × 10−4 mole. (13.76)
The reported coupling to the pellet is only 1.7% of the 1.8 MJ/pulse, or 30 kJ/pulse.
The amount of energy necessary to raise the temperature of the pellet to 108 K is
much greater:
Q = Nmole c v T = (3 × 10−4 mole)(3 × 8.3 J/K mole)(108 K) = 750 kJ,
(13.77)
P1: GFZ
with specific heat c v = 3 Rgas . Lawrence Livermore National Laboratory states that
the pellet “can produce 600 to 1000 times more energy than is put into it.” This
requires the pellet to develop some 600 × 750 kJ = 500 MJ yield, about 10 higher
than our very approximate estimate, which did not include the energy needed to
make the laser beams.
13.10.3 Z-Machine Pinch

In April 2003, Sandia National Laboratory announced that its Z accelerator pro-
duced 0.02 J in 0.1 m/s from fusion. The Z machine used a 20 million ampere,
100-ns pulse through 200 vertical tungsten wires. The tremendous J × B Lorentz
pinch force on the tungsten wires implodes and vaporizes them. The tungsten va-
por is propelled toward the center of the configuration, where it collides with a
6-mm diameter hydrocarbon foam. The resultant shock wave in the foam gener-
ates 200 trillion watts of x-rays, which hit the 2-mm deuterium capsule for fusion.
This tour de force is a long way from commercialization.
13.10.4 Explosion Power

The 94th Congress (1975–76) received testimony on Los Alamos’s Project Pacer,
which proposed using heat from underground peaceful nuclear explosions to gen-
erate electricity. The explosions would be confined to 20-kton weapons since 500-
kton hydrogen bomb explosions could not be contained easily at the Nevada Test
Site. Assuming 10% efficiency of dry geothermal power (Section 13.8) is valid for
Project Pacer, the number of kWh produced by a 20-kton explosion would be
NkWh = 0.1(20 kton)(4.2 × 1012 J/kton)(1 kWh/3.6 × 106 J) = 2.3 × 106 kWh.
(13.78)
At one explosion a day, average electrical power would be
PPacer = (2.3 × 106 kWh/24 h) = 105 kWe = 0.1 GWe . (13.79)
It would take a tremendous number of explosions (4000/day) to satisfy US av-
erage US power of 400 GWe . The US stockpile of 10,000 weapons (2003) would last
only two days at this rate. If a primary costs $1 million, the cost of electricity would
be 10 times its present cost:
($106 )/(2.3 × 106 kWh) = 50 c /kWh, (13.80)
Along the same lines, the Chetek Corporation was formed during the last days
of the Soviet Union in (1991) to commercially explode surplus nuclear weapons
at Novoya Zemela. The purpose of Chetek was not to develop electrical power
but to blow up chemical weapons and destroy surplus bombs. The Foreign Min-
istry opposed this project, as did an American delegation. On the other hand the
Minister of Atomic Power and Industry, Victor Mikhalov, was enthusiastic about
Chetek. Most Americans were concerned about legitimizing peaceful use of nuclear
P1: GFZ
Problems 339
weapons, as well as raising environmental concerns for the Arctic Ocean. These are
reasonable concerns, but they should be compared with Russia’s ability to dispose
of its excess plutonium.
Problems
13.1 Photovoltaic panel area. What is the average electrical power used in your
home based on your electrical bill? Estimate peak power during busy times.
Obtain a cost estimate to power your home with PV panels. What will be the
cost of 1 kWh in summer and in winter with this approach?
13.2 Solar-thermal area. Calculate the land area needed for a full time 1-GWe solar
thermal plant based on the following: 50% collector efficiency, 15% generation
efficiency, 5% storage loss, and average flux of 6 kWt h/m2 day.
13.3 PV costs. What are summer and winter costs of electricity from a seasonally
adjusted collector at the equator and at 45◦ N latitude from PVs with a load
factor of 20%, costs of $2, $5 and $8 per Wpeak with 10%/year capital charges.
13.4 PV storage. What is the cost to store 1 kWh for 1 day under the following
assumptions: A 12-V lead acid battery draws 25 A for 160 min before it drops
below 10.5 volts when it is not used. The battery costs $100, lasts 5 years and
is 75% efficient.
13.5 PV concentrators. Use a ray diagram to show that Winston concentrators
(deep paraboloids) can enhance solar flux by a factor of 10. What concentrator
capital cost cuts the 25 c/kWh cost of PV electricity in half if 75% of the cost
is in semiconductors?
13.6 PV efficiency. The semiconductor SLOium has a band gap of 1.3 eV. Assume
light has a symmetrical triangle distribution beginning at 0.4 μ and ending
at 0.8 μ, and that 80% of the incident light is absorbed in the PV. (a) What is
the efficiency of SLOium? (b) How large a SLOium PV is needed to obtain
1 kWe at peak solar flux? (c) What is peak cogenerative power available from
a 1 kW SLOium PV?
13.7 Photosynthesis. Plants can produce 1 kg/m2 biomass a year. What is the
efficiency of photosynthesis, assuming each atom in sugar releases 10-eV
when burned?
13.8 Biomass. (a) In the 1970s it was estimated that California could annually
produce 1.8 quads biomass energy from 0.5 quad as derived from wastes,
0.6 quads from energy farms and 0.7 quads from kelp. If biomass conversion
was 1%, how much land would need to be dedicated to energy farms to
obtain the quoted energy? (b) Estimate biopower (quads/year, watts) that
could be obtained from manure from the US’s 100 million cows, 60 million
pigs, 7.5 billion chickens, and 290 million turkeys.
13.9 Hydro versus OTEC. (a) How many energy (kilowatthour) can be obtained
from dropping a mass of 1000 kg a distance of 100 m at 90% efficiency? (b)
How many kilowatthour are obtained from cooling 1000 kg of 3◦ C water at
2% efficiency?
P1: GFZ
13.10 Hydro power. How much energy (kilowatthour) are generated by draining
a lake with dimensions 2 km × 8 km × 0.1 km through a height of 0.5 km
at 90% efficiency? What electrical power is produced if the lake was drained
over 10 h? What is the cost filling the lake at 4 c/kWh?
13.11 Pumped storage. (a) Electricity from a 1-GWe plant is used to pump water
200 m upwards to a lake. How much water is pumped over a 12-h night at
90% efficiency? (b) How much energy is recovered during the 12-h day, at
90% efficiency? (c) What is pumped storage power over 12 h of daylight? (d)
What investment is needed to give electricity at 6 c/kWh?
13.12 Wind load factor. What is wind’s capacity factor if global wind produced 37
billion kWh in 2000 from 18.5 GWe ?
13.13 Windmill sizing. Estimate the size of a windmill to sustain a family of four
at a constant 4 kWe . Assume the wind velocity is 6 m/s 50% of the time
and 3 m/s 50% of the time with a windmill efficiency of 25%. Determine the
number of 2 kWe h storage batteries needed to do this if the batteries operate
at 75% efficiency.
13.14 Parafoil windmills. Australians have flown windmills on parafoils at high
altitudes to take advantage of the jet stream. What is the windmill power of
a 5-kWe windmill (at 6 m/s), which is elevated to 2-km height where wind
velocity is 50 m/s in an air density of ρ = ρo e−h/H where h is altitude and H
is 7 km?
13.15 Wind distribution. An Enterek windmill is rated at 1.5 kW in a 9 m/s wind.
How big is the windmill if it utilizes 25% of the wind energy at 9 m/s?
How many kilowatthour can it generate in a month if wind distribution and
efficiency are as follows: Calm wind, 23.8%, η = 0%; 1–2 m/s, 14.2%, 10%;
2–5 m/s, 36.9%, 20%; 5–8 m/s, 13.0%, 30%; 8–13 m/s, 11.2%, 25%; > 13 m/s,
0.9%, 20%.
13.16 Wind conversion efficiency. Show that the maximum conversion efficiency of
a windmill is 16/27 (the Betz limit), when vexit =venter /3. Use mass continuity
(v× Area = constant) and Newton’s second law.
13.17 Passamaquoddy tides. (a) What is stored energy of this 260-km2 bay on
the Maine coast with tidal range of 5.5 m? (Seawater has a specific gravity of
1.025.) (b) Reversible turbines utilize both high and low tides. What is average
power developed at η = 90% from two high and two low tides in 25 h?
13.18 Geothermal. Pressurized water at 300◦ C is used to drive a thermal plant
with 40◦ C output temperature. What is the output power of the plant for a
500-kg/s flow at 50% Carnot efficiency?
13.19 Solar core. (a) What is the kinetic energy of hydrogen nuclei at the sun’s
central temperature of 20–40 million K? (b) What is the electrostatic repulsive
energy of two hydrogen nuclei separated by 8 fermi (8 × 10−15 m)? (c) What
temperature is needed for two deuterium nuclei with average energy to
touch without quantum tunneling?
13.20 DD fusion. What is the fusion power from a 1-m3 deuteron plasma with
a density 1018 particles/m3 at 20 million K if D + D = 4 H + 3.7 MeV and
σ v = 10−24 m3 /s.
P1: GFZ
Bibliography 341
13.21 Cold and bubble fusion. Neither of these ideas was successful. In cold
fusion, the quantum mechanical, WKB transmission probability is e−α
where α is the integral of kdr (k as momentum/h/2π ) between outer
radius r1 , where deuteron thermal energy equals DT repulsive energy, and
inner radius r2 (the sum of DT radii tails of 7 fermi). (a) Set up integrals
to determine the ratio of DT transmission probability at 20 million K to
that at 300 K. (b) Determine the approximate ratio. (c) Bubble fusion: How
much do sono-luminescent bubbles of 1-mm diameter have to collapse to
attain a temperature of 10 million K in order to realize tabletop fusion from
deuterated acetone? How would you detect that bubble fusion took place?
Bibliography
Baldwin, S. (2002). Renewable energy: Progress and prospects, Phys. Today 55(4), 62–67.
Cook, J., J. Beyea and K. Keeler (1991). Potential impacts of biomass production in the United
States on biological diversity, Ann. Rev. Energy Environ. 22, 401.
de Winter, F. (Ed.) (1990). Solar Collectors, Energy Storage and Materials, MIT Press, Cambridge,
MA.
Dracker, R. and P. Delaquil (1996). Progress in solar-electric power systems, Ann. Rev. Energy
Environ. 21, 371–402.
Emanuel, A. and J. McNeill (1997). Electric power quality, Ann. Rev. Energy Environ. 22,
263–304.
Geller, H. (1985). Ethanol fuel from sugar cane in Brazil, Ann. Rev. Energy Environ. 10, 135–164.
Gregory, D. and J. Pangborn (1977). Hydroelectric energy, Ann. Rev. Energy Environ. 2, 279–
310.
Johansson, T., et al (1992). Renewable Energy: Sources for Fuels and Electricity, Island Press,
Washington, DC.
Larson, E. (1993). Technology for electricity and fuels from biomass, Ann. Rev. Energy Environ.
18, 567–630.
Marsden, S. (1983). Methanol as a viable energy source in today’s world, Ann. Rev. Energy
Environ. 8, 333–354.
Masters, G.M. (2004). Renewable and Efficient Electric Power Systems, Wiley, New York.
Mazer, J. (1996). Solar Cells: An introduction to crystalline photovoltaic technology, Kluwer ,
Boston, MA.
McGowan, J. and S. Connors (2000). Windpower: A turn of century review, Ann. Rev. Energy
Environ. 25, 147–197.
Merriam, M. (1978). Wind, wave and tides, Ann. Rev. Energy Environ. 3, 29–56.
Mock, J., J. Tester and P. Wright (1997). Geothermal energy from the Earth, Ann. Rev. Energy
Environ. 22, 305–356.
National Research Council (1991). Potential Applications of Concentrated Solar Photons, Na-
tional Academy Press, Washington, DC.
——— (2001). Energy Research at DOE: Was It Worth It, 1978-2000?, National Academy Press,
Washington, DC.
——— (2001). Laying the Foundation for Space Solar Power, National Academy Press,
Washington, DC.
Norton, B. (1992). Solar Energy Thermal Technology, Springer-Verlag, New York.
P1: GFZ
Office Technology Assessment (1978). Application of Solar Technology to Today’s Energy Needs,
——— (1980). Energy from Biological Processes, OTA, Washington, DC.
Pimentel, D. and E. Terhune (1976). Energy and Food, Ann. Rev. Energy Environ. 1, 171–196.
Renewable Energy World (2001). Renewable Energy Review 4, July 2001.
Sorensen, B. (1995). History and recent progress in wind energy utilization, Ann. Rev. Energy
Environ. 20, 387–424.
US Energy Information Administration (1999). Renewable Energy Issues and Trends, EIA, Wash-
ington, DC.
Wyman, C. (1999). Biomass ethanol: Technical progress, opportunities and challenges, Ann.
Rev. Energy Environ.. 24, 189–226.
P1: GFZ
14
Enhanced End-Use Efficiency
This chapter is about energy success stories and potential success stories. Since the
oil embargo, the United States has reduced its energy-use growth rate from 4.4%
per year (1960–70) to about 1% per year. The nation’s appetite for energy rose from
74 quads in 1973 to 100 quads in 2004, a much smaller rise than the 1972 Atomic
Energy Commission projection of 160 quads for 2000. Electric power consumption
actually grew by 2% per year in the 1990s, reaching an average power of 450 GWe in
2004. This growth was also well below the 1972 Atomic Energy Commission (AEC)
projection of 2000 GWe for the year 2000. The reason energy use by 2000 had not
matched expectations is found in enhanced end-use efficiency. The United States has
saved 50% of energy use on new autos (other than SUVs), houses and refrigerators
since the oil embargo of 1973–74. Appliance standards saved the building of 50
large power plants, which would have consumed 3 quads/year. Energy demand
could be cut by another 50% on new cars and houses, but it must be shown that
these yeoman technologies are cost effective. Over a 10–20 year period, thicker
insulation is cost effective, but it is far cheaper as installed on new construction,
compared to when it is retrofitted to existing houses.
Energy conservation can be accomplished by doing tasks more efficiently, or by
doing without. As consumers sat in long gas lines in the 1970s, President Jimmy
Carter cautioned that Americans were dealing with the “moral equivalent of war”
and that society was in a “malaise.” Carter’s TV appearance in a cardigan sweater
hinted at American’s personal responsibility that they were, indeed, part of the
problem. This was something the electorate did not like to hear. Political willpower
on energy conservation slowed when gas lines disappeared. The drop-off became
clear when Senator Charles Percy’s amendment to add 1 c/gallon for energy re-
search and development was resoundingly defeated. Since those days, the phrase
“conservation of energy” is used less in energy circles because it implies sacrifice,
which is an unpopular concept. Thus, today’s energy policies avoid dictating self-
restraint; instead, they are based on the idea that familiar tasks can be done more
efficiently.
343
P1: GFZ
344 14. Enhanced End-Use Efficiency
14.1 Heat/Cold Storage in Large Buildings

14.1.1 Swedish Design
Many Swedish office buildings need very little heat because they use excess daytime
heat for nighttime warmth, storing energy in concrete floor and ceiling slabs. To
dampen noise, concrete flooring is usually padded, an approach that prevents good
thermal contact between the concrete and the inside air. However, Swedish Ther-
modeck buildings transfer energy to the interior of the concrete through hollow-
core tubes in floor slabs. The tubes are cast into the concrete slabs to reduce slab
weight while minimizing mechanical deflections, similiar to I-beams. Although
Stockholm with 3580 ◦ C day/year (6444 ◦ F day/year, Section 11.3) is colder than
Chicago, Thermodeck buildings use only 4 kWh/ft2 of electric resistance heating
per year, a cost so low that it is not necessary to hook Thermodecks to buy from
Stockholm’s district heating system. This system also works well in the summer,
using night air to precool buildings.
14.1.2 Size and Temperature

We begin with energy gains/losses to see if it is possible to operate Thermod-
eck buildings with essentially no added heat. A single-occupant office is 2.4-m
wide by 4.2-m deep by 2.7-m high, for a 10-m2 area and 27-m3 volume. A cold
day in Stockholm of –9◦ C (16◦ F) has an inside-to-outside temperature difference
T = 21◦ C − (−9◦ C) = 30◦ C (54◦ F). A person’s body heat produces 100 W, and
lights and machines in an office contribute an additional 300 W. Because Swedish
offices must have windows, Thermodeck buildings have a large surface-to-volume
ratio, increasing insulation needs. Each office has 1.5 m2 of triple-glazed windows
(USI = 2) and 5 m2 of wall surface (USI = 0.25). Natural infiltration enters in a room
at 5 m3 /h during unoccupied hours. During the occupied hours infiltration is in-
creased 20 m3 /h to ensure good air quality. Solar gain is a modest 30 W/room
during winter occupied hours.
The heat-transfer rate from infiltration is
d Q/dt = ρ(d V/dt)cT, (14.1)
where air density ρ = 1.3 kg/m3 , air flow rate is dV/dt, air specific heat c =
1000 J/kg ◦ C and T is the inside-to-outside temperature difference. The loss rate
during the occupied hours, when dV/dt = 5 m3 /h = 0.0056 m3 /s, is
d Q/dt = (1.3 kg/m3 )(0.0056 m3 /s)(1000 J/kg ◦ C)(−30 ◦ C) = −218 W. (14.2)

The infiltration loss with dV/dt = 5 m3 /h during unoccupied hours is 54 W. Heat
transfer rates per room are shown in Table 14.1:
Heat is transferred in the morning to raise the temperature above To from the
evening hours. The temperature T of the room air is
T = To + (d Q/dt)t/C, (14.3)
P1: GFZ
14.1. Heat/Cold Storage in Large Buildings 345
Table 14.1. Heat transfer in a Thermodeck room.

Source of heat gain/loss per room (watts) day night
Infiltration –218 –54
Wall, UAT = (0.25)(5)(30) –38 –38
Window, UAT = (2)(1.5)(30) –90 –90
Total loss –346 –182
Gain: solar (30), person (100), equipment (300) 430 0
Net = gain - loss +84 –182
where C is the slab heat capacity/m2 and dQ/dt is internal heat rate in W/m2 . The
heat capacity of the 30-cm thick slabs is C = 120 Wh/m2 ◦ C, which includes an
extra 20% to account for the heat capacity of walls and furnishings. Using these
values, the time dependence of the room temperature is
occupied (8 W/m2 ) T = To + 0.08t (14.4)
unoccupied (–18 W/m2 ) T = T1 − 0.18t, (14.5)
with t in hours. During a 10-h workday room temperature remains fairly constant,
rising by 1◦ C (0.08 × 10). Overnight room temperature drops by about 2◦ C (–0.18 ×
14), a drop that is lessened by adding a little heating to the room. See Fig. 14.1.
14.1.3 District Heating

Sweden has over 80 district heating systems that supply heat to downtown areas.
These systems use industrial waste heat or less expensive fuels. Swedish systems
use hot water at rates as large as 300 MWt . Water is pumped through 100 km of
Figure 14.1. Thermodeck in

summer. Outdoor 12◦ C night air
is transferred to the hollow cores
in the concrete slab floor, to cool
the building by day when it
reaches 30◦ C outside. The 18◦ C
spread in outside temperature
gave only a 1◦ C spread in room
temperature (A. Rosenfeld in
Hafemeister (1985), p. 155).
P1: GFZ
pipes, in a process that is quite dependable. US district heating systems, on the other
hand, often fail because they usually use high-pressure steam, which is harder to
contain. The energy cost of steam leaks can cost as much as $1 million/year:
“. . . . US city-scale district heating systems owned by utilities have, up until now,

enjoyed little success when compared to European systems, primarily because
European systems use hot water instead of steam.”
[Industrial and Commercial Cogeneration, OTA, 1983]
14.1.4 Thermal Storage to Reduce Peak Power

Daytime air conditioning could be significantly reduced by taking advantage of
the night’s coolth and the fact that power is cheaper at night. During the deep air
conditioning season, chillers are run at night, when electricity is cheaper, to precool
slabs. This approach also takes advantage of improved efficiencies from lower night
temperatures. Implementing off-peak electricity usage may not give considerable
energy in terms of kilowatthours, but the tactic does save on the expense of daytime
peak power charges; some $100/kW year is saved by operating thermal storage
systems at night. Finally, this approach downsizes cooling systems by allowing
them to be used on a 24-h basis.
Internal heat gains are a dominate feature of large buildings, a fact that makes
air conditioning a necessity in them. This excess demand causes severe daytime
summer peak power loads, as much as 2–3 times the nighttime load. The fraction of
US homes with air conditioning reached 77% by 2001, increasing US peak demand
by 2 GWe /year. The 5%/year growth rate for new commercial buildings (replace-
ment plus growth) adds 2.5 billion ft2 /year of floor space, driving peak demand
by 1.6 GWe /year. Residential and commercial air conditioning each used 80 GWe ,
for a total peak load of 160 GWe . This could be reduced with off-peak cooling with
thermal storage.
In 1986 engineers realized that Stanford University’s daytime cooling needs were
going to rise from 5 MW to 8 MW in a decade. Their first response was to consider
adding a megawatt of chillers and cooling towers at a cost of $1.5 million. However,
Stanford discovered that $1 million could build a 1 million gal insulated water
tank for cold storage, one that would be connected to existing chillers. Since this
discovery, Stanford has shifted from cold water storage to ice storage. Stanford
now makes 4000 tons of ice per night, thus running the third largest ice production
facility in the world. By 2003 Stanford met one-third of its 25-MWe peak power
needs with ice storage, saving $500,000/year.
The headquarters of the Alabama Power Company in Birmingham, Alabama,
installed five large ice cells to contain 550 tons of ice for cooling for its 110,000-m2
building. Ice storage on the basis of floor-area density basis is 5 kg per 1 m2 floor
space (550 tonne/110,000 m2 ), with a stored energy density of
Q/m2 = mL fusion = (5 kg/m2 )(3.4 × 105 J/kg) = 1.7 MJ/m2 , (14.6)

P1: GFZ
14.2. Improved Lighting 347
where L ice is the latent heat to freeze ice. The electrical power/m2 needed to make
ice during 16 off-peak hours is, thus,
P/m2 = (Q/m2 )/(COP)(t) = (1.7 MJ/m2 )(1 kWh/3.6 MJ)/(2.5)(16 h) = 12 We /m2 .
(14.7)
(COP is coefficient of performance, in this case 2.5.) The total power required for ice
storage is, therefore,
P = (12 We /m2 )(1.1 × 105 m2 ) = 1.3 MWe . (14.8)
Without thermal storage it would take 2.8 MWe to cool the building, more than
twice the 1.3 MWe used with ice storage. Coolth stored in ice supplies 2/3 of a
day’s cooling requirement with 1/3 coming from direct daytime cooling. Since ice
storage covers 2/3 of daily summer heat gain, the total daily heat gain in summer is
(1.7 MJ/m2 )/(2/3) = 2.6 MJ/m2 . This represents an average heat input power/m2
over 8 h of daytime of
P/m2 = (Q/m2 )/t = (2.6 MJ/m2 )/(8 h)(3600 s/h) = 90 W/m2 . (14.9)
2
Thus, thermal storage supplies about twice the 43-W/m gain of Thermodeck.
14.2 Improved Lighting

Lighting is of great interest to physicists since it contains a great deal of basic atomic
and plasma physics. US lighting consumes 450 billion kWh/year. This amounts to
50 GWe average power, or 12% of US average electrical power. About half is used
for incandescent lighting and the other half is predominantly used for fluorescent
lighting. The units used in lighting are as follows:
r 1 lumen = 1/673 W of visible light
r 1 lux = 1 lumen/m2 of illuminance
r 1 foot candle = 1 lumen/ft2 = 0.0929 lux.
14.2.1 Efficacy and Efficiency

Technical advances since the oil embargo have greatly increased efficacy (lumens/
watt) of lighting. The incandescent bulb (invented in 1879 by Thomas Edison) has
an efficacy of 17 lumens/watt for the 100-W bulb, while the cooler 40-W bulb
has a smaller efficacy of 12 lumens/watt. Mercury fluorescent lighting (invented
in 1937) has an efficacy of 80 lumen/watt but improvements have raised it to 120
lumens/watt. Compact fluorescent bulbs have an efficacy of 60–90 lumens/watt. In
commercial or industrial situations, sulfur lamps have 10 times greater efficacy than
mercury fluorescents, but they are limited to large-scale applications. Halogen floor
lamps produce considerable light, but their high thermal power has a propensity
for starting fires. New lamps using compact fluorescent bulbs have five times the
efficacy of halogen lamps and they are far safer.
P1: GFZ
The lighting efficiency of a 100-W bulb is given by
η = (17 lumen/watt)/(673 lumen/watt) = 2.5%, (14.10)
while fluorescent tubes with an efficacy of 120 lumens/watt are 18% efficient
(120/673). Industry prefers fluorescent tubes because of energy savings, but even
more because the maintenance of one is much less labor-intensive than for incan-
descent bulbs. Fluorescent tubes last 10,000 h, 10 times longer than incandescents,
requiring only one-tenth as many bulb changes. Light-emitting diodes are 1.5 times
more efficient than fluorescent lamps, but their higher cost limits their use to ad-
vertising signs and very small lights. Research is underway to lower these costs.
Mercury fluorescent tubes convert 2/3 of input electrical-discharge energy into
ultraviolet photons (λ = 0.254 μ, 4.9 eV). The radiation diffuses through the tube of
plasma, being absorbed and reemitted by other mercury atoms until it is converted
to the visible range (λ = 0.5 μ, 2.2 eV) by phosphors on the tube wall. Electromag-
netic core-coil ballasts provide starting voltages for the plasma discharge, but the
process consumes 25% of input energy. High-frequency, solid-state ballasts, oper-
ating at 30 kHz, reduce these losses to raise efficacy from 80 to 90 lumens/watt.
The ballasts can also be used to dim the fluorescents.
14.2.2 Compact Fluorescent Lights

Fifteen-watt compact fluorescent lights (CFL) produce the equivalent light of 60-W
incandescent bulbs, saving 75% of lighting energy. CFLs cost $2 per bulb as com-
pared to an equivalent lifetime (10,000 h) of incandescent bulbs ($7 for 10 bulbs).
The cost of electricity for continuously operating ten 60-W bulbs 24 h/day over
10,000 h (14 months) is
(0.06 kW)(10, 000 h)($0.10/kWh) = $60, (14.11)
for a total cost of $67 (bulb cost plus electricity cost). On the other hand, the cost of
operating the 15-W CFL is $15, plus “$2 purchase cost” for a total of $17. The
monthly savings works out to $50/14 months = $3/month. At this rate it takes just
a month to pay for a CFL. However, if the bulb is used only 4 h a day, the economics
is figured over a longer period.
Each CFL bulb saves 45 W over its 10,000-h lifetime, for a total of (0.045 kW)(10,000
h) = 450 kWh per CFC. Since it takes a little less than a pound of coal to make a
kWh, this corresponds to a savings of 200 kg coal per CFC bulb. If all 100 million
US housing units each used 1 CFL for 4 h a day, the savings would be
(100 million CFL)(0.045 kW)(8766 h/6 year) = 6 × 109 kWh/year. (14.12)
This is equivalent to the annual produced by a 0.7-GWe plant (100% load factor),
saving 3 million tons of coal. The adoption of CFLs is increasing as newer CFCs fit
into existing lamps better than earlier models.
P1: GFZ
14.2. Improved Lighting 349
14.2.3 Isotopic Enhancement

Mercury consists of six predominant isotopes and one minor isotope, 196 Hg. Each
Hg isotope has an emission spectrum that is slightly different from the other Hg
isotopes. This difference is caused by the isotope shift, which results from the
small electrostatic energy differences between the variable-sized isotopic nuclei
and the inner s-electrons. Efficacy of fluorescents can be increased by adding
196
Hg, a process that makes new photons at slightly different energies. This al-
lows more photons to arrive at the tube wall, thus reducing absorption in other
pathways.
To determine if addition of 196 Hg could be effective, the Hg isotope shift is com-
pared with the Doppler width of a resonance. A mercury isotope shift of 5 GHz
corresponds to an energy shift
E = h f = (4.1 × 10−15 eV s)(5 × 109 Hz) = 2.1 × 10−5 eV, (14.13)
where h is Plank’s constant. The atomic lines are broadened by motion of Hg ions.
The Doppler shift of a moving Hg atoms at 900 K (kT = 0.075 eV) is
E = (v/c)E = (1.5kT/mc2 )1/2 E = (1.5 × 0.075 eV/200 × 109 eV)1/2 (4.9 eV)
= 3.8 × 10−6 eV. (14.14)
Since the isotope shift is six times larger than the Doppler width, thermal broad-
ening should not prohibitively overlap the isotopic lines. By adding the seventh
isotope, 196 Hg, the number of energy pathways in the plasma is increased by about
one-seventh, reducing the number of nonradiative de-excitations by about one-
seventh. This raises efficacy by 15%, or about 10 lumens/watt. This approach is
technically feasible, but the economics is not yet favorable.
14.2.4 Phosphors
Mercury’s 4.9-eV (0.254-μ) ultraviolet radiation is twice the energy of visible light.
It should be possible to find phosphors that produce two visible photons from
one ultra violet photon. A quantum-efficiency of 200% doubles lighting efficacy
and could save 20 GWe . It was discovered (1999) that the phosphor LiGdF4 :Eu3+
produced two photons for each ion in a xenon tube. Since the 7.2-eV xenon transition
is more energetic than mercury’s 4.9-eV transition, this approach is not feasible in
fluorescent tubes.
14.2.5 External Magnetic and Electric Fields

Magnetic fields create new atomic lines by hyperfine splitting of mercury spectra,
giving new energy pathways that allow photons to diffuse toward the tube walls
with fewer of nonradiative deexcitations. This approach at B = 0.06 T increased
efficacy by 15%. Another approach uses electric fields. Mercury plasmas, excited
with 500-MHz electric waves on the inner surface of a fluorescent tube, creates an
P1: GFZ
E-field that confines the plasma to a region closer to the wall, thus, reducing the
volume of the tube in which losses due to collisions can take place. This approach
was shown to increase efficacy by 30%.
14.2.6 Sulfur Light-Guide Systems

High-lumen, high-efficiency sulfur lamps show great promise. In a 1995 demon-
stration they were connected to hollow light guides at the Forrestal Building in
Washington, DC. The original lighting used 140 kW to illuminate the Forrestal’s
outdoor plaza with 280 500-W incandescent bulbs. Fluorescent bulbs were installed,
dropping power by 65% to 49 kW with 280 175-W tubes. The fluorescent tubes were
replaced with two 5.9-kW sulfur lamps without electrodes, saving 90% of the origi-
nal electrical power. The sulfur lamps were placed at the ends of a 73-m light guide,
which internally reflected and delivered light along the guide path. The sulfur lamp
illuminance was 137 lux (lumens/m2 ), as compared to 44 lux with mercury lamps.
However, the sulfur illuminance varied from 100 to 500 lux, while the many flu-
orescent lamps were much more constant. In a similar fashion, the Smithsonian
Museum reduced its power from 23 kW to 18 kW while tripling average illumi-
nance. Sulfur lamps with light guides are limited to high output situations for
industry and the outdoors.
14.2.7 Lighting Standards and Smart Lighting

Lighting in grocery stores can approach the headache level. Lighting standards in
industrial nations vary by a factor of 2–3 for no particular reason. The US classroom
standard rose from 250 lux in 1950 to 1500 lux in 1970, then dropped to 800 lux in
1980, while other industrialized nations have been using about 500 lux in class-
rooms. The use of sensors that control lighting levels based on room occupancy
can save considerable lighting and air conditioning. See Fig. 14.2.
A/C Figure 14.2. Dimmed

Dimmed
lighting in commercial
lighting buildings Electronic ballasts
Lighting A/C allow fluorescent lights to be
dimmed according to room
occupancy and daylight.
Other Other During peak periods, 75% of
lighting and 25% of air
conditioning were saved by
the use of sensors that control
1 3 5 7 9 11 13 15 17 19 21 23 1 3 5 7 9 11 13 15 17 19 21 23 light levels. (Adapted from A.
Time of Day Time of Day Rosenfeld, LBL)
P1: GFZ
14.3. Improved Windows 351
14.3 Improved Windows

The annual energy lost through windows in 100 million US residences (Section
11.2) is
dQ/dt = (# of homes)(window area/home)(U A)(◦ C day/year)(s/day)
= (100 M)(15 m2 )(6 W/m2 K)(2800◦ C day/year)(86,400 s/day)
= 2 × 1018 J/year, (14.15)
or 2 quads/year. Window loss in commercial and industrial buildings is similar,
raising the total to 4 quads/year, or 4% of US energy use. This loss can be mitigated
with new technologies, which we review here.
14.3.1 Spectral Selective Glazing

Glass has the marvelous property that it acts like a one-way diode. Glass transmits
visible light from the 6000-K sun into buildings, while absorbing infrared that is
directed outward from a building’s 300-K interior. As improving wall insulation
lessens the flow of energy through walls, improving windows lessens the loss of IR
heat, which accounts for 2/3 of heat loss through double-glazed windows. Low-
emissivity (low-E) coatings reduce winter heat losses by reflecting 90% of incident
thermal IR (λ ≈ 10 μ) back into rooms. Conversely, traditional glass with its high
emissivity absorbs 90% of interior’s IR flux, eventually conducting much of it to
the outside.
Convection is the dominant heat path for low-E coatings, constraining U-values
to 30% of traditional glass. Low-E coatings are also useful in summer, as they
reflect away IR emitted by warm surfaces outside the house and near-IR from the
sun. Thus, low-E glass reduces heating in winter and air conditioning in summer.
Switching to low-E windows can reduce energy costs in typical Boston homes from
$900/year to $600/year. Window manufacturers started making low-E windows
in 1991, capturing 35% of the US market by 1996.
Low-E windows use materials selected to transmit visible light and reflect, rather
than absorb, infrared. Wide-band tin and indium oxide semiconductors can be
doped to produce IR reflection. IR reflection can also be done with very thin layers
(10–20 nm) of silver (Fig. 14.3). Reflectivity of low-E windows is based on plasma
physics. Electromagnetic waves above the plasma frequency are transmitted, while
EM waves below the plasma frequency are reflected. An example of this is the way
radio transmission is better at night. Because darkness lacks ultraviolet, the night-
time ionosphere has fewer free electrons at 1000-km altitude, giving a lower plasma
frequency of about 3 MHz. Since radio waves have a frequency under 3 MHz, they
are reflected from the ionosphere, giving signals from 10,000 km away. In a sim-
ilar fashion, the doping level in low-E semiconductors determines free-electron
density, analogous to how UV changes free-electron levels in the ionosphere. By
increasing doping levels, plasma frequencies can be increased. The low-E materi-
als SnO2 :F and In2 O3 :Sn reflect infrared at wavelengths longer than 1.5 μ, while
P1: GFZ
Figure 14.3. Low-E window transmittance (T) and reflectance (R). Dotted curves are for
a single-layer oxide, and solid curves are for a double thin metal layer. Note that higher
transmittance accompanies lower reflectance (Rosenfeld et al., 2000).
transmitting visible light and near IR at wavelengths under 1.5 μ. To obtain the 1.5-
μ transition, an electron density of 5 × 1026 /m3 is chosen to give plasma frequency
(λ = c/ f = 2πc/ω)
ωp = (ne2 /εo m) = 1.3 × 1015 rad/s, (14.16)
14.3.2 Aerogel
Using many layers of glass to greatly reduce energy loss is expensive and unaes-
thetic. A better way to make high-R glass is to use aerogel, making windows as
thermal resistant as wall insulation (RSI = 2, REng = 10). Aerogel consists of tiny,
10-nm diameter silica particles in a honeycomb structure. The poor thermal contact
between these particles gives large thermal resistance. In addition, the 10-nm pores
are much smaller than the mean free path of an air molecule. The inability of air to
easily traverse the aerogel layer greatly reduces convection losses. Containing aero-
gel in a vacuum at 10% of an atmosphere further reduces convection losses. Aerogel
must be contained inside double-glazing since it is sensitive to mechanical impact.
14.3.3 Smart Windows

Electronic (electrochromic) windows act as electronic blinds, switching from light
transmission to light reflection with the application of a few volts. A manual switch
P1: GFZ
14.4. Heat Pumps 353
or a motion detector allows sunlight to enter when rooms are occupied and blocks
sunlight to conserve air conditioning when rooms are empty. One smart-window
approach uses pulsed-laser deposition of LiNiO2 powder onto glass. Optical trans-
mission is 70% in the center of the visible (0.55 μ) for 150-nm films. These windows
can save peak power use by 20–30% in many commercial buildings. Another ap-
proach uses liquid crystal molecules, dispersed in a polymer matrix. With no ap-
plied voltage, the random orientation of molecules makes the window dark as the
molecules scatter light. When an electric field is applied the molecules align to give
an index or refraction equal to that of the matrix and 80% of light is transmitted. For
this to happen electrochromic polymers must be cheaper than their present cost of
$1000/m2 .
14.3.4 Daylighting with Shelves and Pipes

Beyond energy savings, a benefit of smart windows and other daylighting tech-
nologies is that people appear to function better in natural light than in artificial
light. One study claims this result by observing higher SAT scores than others tak-
ing the test, while another study claims harder work when using daylight rather
than artificial lighting. Work level has been associated with the nature of workplace
light. A New Yorker cartoon commented that “dim fluorescent lighting is meant to
emphasize a general absence of hope.”
Physicists have long used light pipes to bring photons from NaI scintillators to
photomultiplier tubes. This idea has been adapted to light shelves, which reflect
sunlight onto ceilings, producing diffuse scattered light. Intensity of reflected light
is adjusted by rotating the angle of the shelf. Similarly, light pipes can transmit light
to far corners of buildings. This is accomplished through the spreading of light
along a pipe through its translucent bottom. More recently, insulation made up of
small-diameter transparent straws allows solar gain, while retaining reasonably
high R-values.
14.4 Heat Pumps

A heat engine transfers heat from a region at a hotter temperature TH to one at
a colder temperature TC , while performing work derived from the difference in
heat energy (Q = QH − QC = W). On the other hand, a heat pump uses external
work to transfer heat from a colder source to a hotter sink. The heat deposited
in the high temperature sink is the sum of the absorbed low temperature heat
and the spent work. Refrigerators cool iceboxes and warm buildings, while air
conditioners cool buildings, while heating the outside. Another task for a heat
pump is to heat buildings, while cooling the ground. This is accomplished by the
geothermal heat pump, which relies on the principle that cooler heat energy coming
from the ground can be warmed with external work. About 2 million heat pumps
of all types are bought annually. Here we discuss vapor compression heat pumps,
rather than absorption heat pumps.
P1: GFZ
The efficiency of a heat engine is the ratio of net output work divided by input
heat at a higher temperature, η = Wnet /QH . The coefficient of performance (COP)
of a refrigerator or heat pump is the converse, as it is the ratio of transfer of useful
heat divided by the net input work, or COP = Quseful /Wnet . Carnot cycle COPs for
cooling and heating heat pumps are:
COPcool = QC /W = 1/(QH /QC − 1) = 1/(TH /TC − 1) = TC /T (14.17)

COPheat = QH /W = 1/(1 − QC /QH ) = 1/(1 − TC /TH ) = TH /T, (14.18)
where T = TH − TC .
A geothermal heat pump that heats a room at 21◦ C (70◦ F) from an outside heat
source at 8◦ C (46◦ F) has a Carnot COP of 22.6. But the Carnot cycle is not appropriate
since isothermal processes are not used and working fluids have phase transitions
in compressors and evaporators. The working fluid follows a path on a pressure-
volume diagram across a mixed liquid–vapor region (water’s “steamdome”) with
phase transitions. The COP of a cycle is based on available energy of enthalpy,
that is the sum of internal energy and the product of pressure and volume (h =
Uinternal + pV). For throttling valves that do no work, h 1 = h 2 . The work and heat
equations shown below use the enthalpy subscripts shown in Fig. 14.4:
W = h4 − h3 (14.19)
QH = h 4 − h 1 (14.20)
QC = h 3 − h 2 (14.21)
The COPs for the two heat pump cycles are as follows:
COPcool = QC /W = (h 3 − h 2 )/(h 4 − h 3 ) (14.22)

COPheat = QH /W = (h 4 − h 1 )/(h 4 − h 3 ). (14.23)
Figure 14.4. Vapor compression cycle for heat pumps. Heat pump cycles move in counter-
clockwise direction as compared to heat engines, which move clockwise. On the pressure
versus enthalpy plot, (1–2) is a constant enthalpy expansion in a throttling valve, (2–3) is
an evaporator in the freezing compartment or in the ground for a geothermal heat pump,
(3–4) is a compressor to raise pressure, and (4–1) is a heat exchanger in the room, where the
working fluid condenses.
P1: GFZ
14.4. Heat Pumps 355
Using enthalpy tables for R-22 refrigerant, a COP of 21.5 is determined for the
cycle described above, slightly less than the Carnot value of 22.6. However, ac-
tual heat pumps do not attain these COPs, as there are many loss mechanisms. A
major cause for a smaller COP is temperature differences in the heat exchanger.
Heat transfer near ground evaporators takes heat away, making the evaporators
colder than normal ground temperature. Similarly, condensers are hotter than the
rooms they service. Oak Ridge National Laboratory scientists measured condenser
fluids at 50◦ C (29◦ C above room temperature) and evaporator fluids at –1◦ C (9◦ C
colder than the ground). Using these real temperatures, the Carnot COP drops
dramatically from 22.6 to 7.3.
Heat pumps sold before 1978 suffered from large heat-exchanger temperature
differences, inefficient compressors, and heat leaks, which cumulatively dropped
the COP to a sluggish 2.2. The COP of current heat pumps is 3.0 to 3.5, but in fact
the actual COP depends on local weather and usage. In summary, measured COPs
are much less than Carnot estimates because (1) friction and heat loss dissipate
energy, (2) heat exchangers are too hot or too cold, and (3) heat pumps do not use
the Carnot cycle, which has isothermal expansions and compression.
14.4.1 Geographic Best Locations

The most cost effective locations for geothermal heat pumps are those with moder-
ate climates, such as the region from San Francisco to Tennessee to Washington, DC,
as opposed to the Dakotas, which are too cold, and the Gulf Coast, which is too hot.
This conclusion can be verified with a quick calculation. We assume actual COP is
10% Carnot COP. A house in a moderate location with 10◦ C outdoor temperature
and 20◦ C indoor temperature has a COP of
COP = TH /10T = 293 K/10(293 K − 283 K) = 3. (14.24)
On the other hand, a house in the far north that uses ice water as a heat source at
0◦ C has a much lower COP:
COP = TH /10T = 293K/10(293 K − 273 K) = 1.3. (14.25)
14.4.2 Second Law Efficiency

The second-law efficiency of a heat pump is the ratio of the measured (actual) COP
to the best possible Carnot COPCarnot between the two temperatures. Alternatively,
second law efficiency is the ratio of the minimum energy to do a task to the actual
energy to do a task. For the heat pump, the second-law efficiency is
ε = E min /E actual = COPactual /COPCarnot = 3/22.6 = 13%. (14.26)
In spite of the fact that the second law efficiency of a heat pump is low, a heat
pump with COP = 3 will heat a house with 1/3 the energy that a furnace consumes
by burning fossils at 100% efficiency. A low second-law efficiency indicates that
there is room for improvement in heat pump or furnace design.
P1: GFZ
This definition of second law efficiency ignores the source of the electricity for the
heat pump. A typical power plant produces electricity with a first law efficiency of
η = 1/3, while a modern combined cycle gas turbine (CCGT, Section 10.6) gets η =
0.60. Since the definition of second-law efficiency is the ratio of minimum possible
energy to the actual energy, we must take into account the source of the electricity.
In running heat pumps, twice as much energy is needed when electricity comes
from traditional power plants instead of CCGTs. Thus, ε should be divided by two,
giving ε = 7%. Second law efficiency does not take into account economics; it only
describes the potential use the technology.
14.4.3 Ice Ponds for Seasonal Storage

On the other hand, we can do better than Carnot if we are clever, using seasonal
storage to bring the coolth of winter into the summer. The cooling value of a ton
of ice varies between $1/ton for air conditioning with cheap power to $5/ton for
process cooling refrigeration. Ice is produced essentially cost free with surface freez-
ing, while sprays in cold Minnesota require 0.5 kWh/ton electricity and warmer
Washington, DC requires 2 kWh/ton. As a test of these ideas, Ted Taylor built a
large demonstration ice pond for the Prudential Insurance Company in New Jersey
(1980s), but it shrank by summer time. Along the same lines, experimenters at Oak
Ridge National Laboratory built a test house in which ice was naturally made in
the basement in winter, then used to cool the house in summer. The Annual Cycle
Energy System house developed 3.1 times the energy actually needed for cooling
the house. However, the payback period was 25 years because of the initial high
cost of $8,000 in 1985 dollars.
14.5 Improved Appliances

14.5.1 Refrigerators
Refrigerators were egregious energy guzzlers before the oil embargo, but now their
consumption has been cut by 75%. The average 1974 refrigerator consumed 1800
kWh/year (Fig. 14.5). The high level was adopted as the starting point standard by
the California Energy Commission in 1977 to give manufacturers time to improve
their products. California tightened the standard to 1500 kWh/year in 1979, to 1000
kWh/year in 1987 and to 700 kWh/year in 1993. The US standard for refrigerators
dropped to 450 kWh/year in 2001. Levels as low as 200 kWh/year have been
obtained with demonstration models.
In the mid-1980s there was a great debate on the setting of energy standards
for US appliances. At that time, the major appliances (refrigerators, freezers, space
and water heaters, air conditioners) consumed 8.9 quads/year, which was 12% of
the national energy use of 74 quads/year. Lawrence Berkeley National Laboratory
estimated that energy conservation measures could save about 30% of this (2.7
quads/year) with an investment of $7 billion/year to recover $17 billion/year.
P1: GFZ
14.5. Improved Appliances 357
Figure 14.5. Electricity use by new US refrigerators, 1947–2001. Heavy line with dark squares
represents sales-weighted average annual kilowatthour use of new refrigerators, adjusted
for increasing volume. Lighter line with open circles represents volume growth, from 8
cubic feet to 20 cubic feet. The right scale shows the number of large (1 GWe ) base-load (5000
h/year) power plants required to power 150 million refrigerators and freezers (kilowatthour
on left scale). The difference between 1974 (1800 kWh/year) and 2001 (450 kWh/year) is 1350
kWh/year. The eventual savings is equivalent to 50 avoided 1-GWe plants (right scale). At
8 c/kWh, the avoided annual cost is $16 billion (Rosenfeld, 1999).
In July 1985, the US Appellate Court ruled that the executive branch was re-
quired to implement mandatory energy standards for these appliances. Prior to
that time the Department of Energy avoided the mandated standards by adopting
the no-standard standard. Judge Robert Bork rejected the no-standard standard order-
ing the use of a reasonable discount rate to determine future benefits. (A life-cycle
cost analysis is carried out in Section 16.4.)
14.5.2 1975 Refrigerator

Before the oil embargo, refrigerators were designed with thin walls having little
insulation. This allowed for greater volumes and smaller initial costs. In the very
old days, refrigerator heat exchangers were placed in domes on top for good heat
exchange, but they were considered unsightly. Now the heat exchanger is placed
on the back of the refrigerator where it ends up next to a wall, reducing heat
transfer. The motors and compressors were placed below the box, where they leak
P1: GFZ
heat upwards into the box, requiring further refrigeration. The 1975 refrigerator
had 57% heat gain through the walls, 23% through interior heaters and fans, and
12% through gasket area. Only 6% of heat gained was from food cooling and door
opening caused only 2%. Automatic-defrost heaters consumed 110 kWh/year and
freezer case heaters, which prevent ice from forming consumed 60 kWh/year. The
motor compressor unit of old had a COP of 0.9. Good engineering and market
forces were not part of this picture.
14.5.3 1995 Refrigerator

New refrigerators that consume 700 kWh/year have these improvements.
r polyurethane foam insulation
r 5 cm of super insulation at RSI 3.3 (REng 19)
r 2.5 cm gas filled panels at RSI 2.5 (REng 14)
r more efficient motors and compressors
r greater condenser area, lowering temperatures from 122◦ F to 101◦ F
r greater evaporator area, lowering temperature from 18◦ F to 9◦ F
r elimination of case heaters (or addition of a switch)
r double gaskets on doors.
14.5.4 Shifting Freezing to the Freezer

Refrigerators typically operate at 4◦ C (40◦ F) while freezers typically operate at –7◦ C
(20◦ F). A thermodynamic penalty is extracted when the colder freezing compart-
ment is placed inside a refrigerator. This forces a larger temperature difference on
the cycle, lowering the COP. We assume for both cases that heat is released from
the condenser coil into room air at 35◦ C (95◦ F). For the case of a refrigerator with
an inside freezer, the Carnot COPfreezer = TC /T = (266 K/42 K) = 6.3. This can
be compared to the case of a refrigerator that has an adjacent compartment, which
has a 30% higher Carnot COPrefrigerator of (277 K/31 K) = 8.9. Refrigerators with
inside freezing compartments have the added problem of bothersome ice, which
raises thermal resistance on the coils.
14.5.5 National Energy Savings

Energy standards reduced annual refrigerator consumption from 1826 kWh/year
to 475 kWh/year. The improvement calculated below saves the US considerable
energy on its stock of 200 million refrigerators and freezers:
kWh = (2×108 )(1826 − 475) kWh/year = 2.7×1011 kWh/year = $27 billion/year,

(14.27)
with 10 c/kWh electricity. A 1-GWe power plant with a 80% capacity factor
P1: GFZ
14.6. House Doctors 359
produces
(8766 h/year)(106 kW)(0.8) = 7.0 × 109 kWh/year. (14.28)
The number of 1-GWe power plants saved is
(2.7 × 1011 kWh/year)/(7.0 × 109 kWh/GWe year) = 40 GWe . (14.29)
Predictions for future US energy savings due to improvements in other appliance
design allow another 20 GWe , for a total of 60 GWe , which is 13% of US grid
production.
14.6 House Doctors

US buildings built before the oil embargo in a time of cheap energy often had
no insulation in warm climates like California, or too little insulation in colder
climes. A US house before the embargo using oil or gas heat consumed 73 million
Btu/year (77 GJ, 12.5 bbl of oil). In 1990 the average new US house consumed 40%
less energy at 43 million Btu (45 GJ, 7.4 bbl). Air infiltration is responsible for 20% of
the total loss, and hot air systems lose another 20% through duct leaks and furnace
inefficiency.
In this section we consider the benefits a house gains from occasionally “seeing”
a doctor, much as a human body needs occasional visits to a doctor. Gautam Dutt
from Princeton University became the nation’s first house doctor when he made
the first serious house calls to buildings in the 1970s. Prior to that time there was
little scientific study on buildings owing to the industry being very decentralized.
The Princeton group discovered that attic bypass paths allowed warm air to travel
around ceiling insulation through channels in walls to cold attics. We followed the
work of Dutt by creating a house doctor laboratory for Cal Poly architecture stu-
dents. The laboratory was named, The Arthur Rosenfeld House Doctor Laboratory, for
the man who has been the main leader in improving energy use in US buildings.
It is obvious that retrofitting an existing building is more complicated and expen-
sive than constructing it correctly in the first place. However, some energy-saving,
retrofit measures can be easily carried out. In the 1980s it was estimated that an
investment of $1000 would save 25% of a home heating bill and $2000 would save
40%. In this section we describe the blower-door which measures infiltration and
locates leaks (Fig. 14.6).
14.6.1 Stack-Induced Infiltration

The pressure at the top of a chimney or attic of height h is reduced with respect to
the ground level by
pstack = −ρgh = −(1.3 kg/m3 )(9.8 m/s2 )h = −13h Pa, (14.30)
with h in meters. If this reduction were the only effect, air would not rise, but re-
main in static equilibrium. But warmed air rises because its density is inversely
P1: GFZ
Figure 14.6. Cal Poly blower door. Jim Woolaway’s blower door measured infiltration area
in author’s home. The difference between inside and outside pressures is measured as a
function of fan speed. Leaks were discovered by over pressurizing the house, then following
smoke emanating from a smoke stick to the leaks. (Photo by Tony Hertz, San Luis Obispo
Tribune)
P1: GFZ
proportional to its absolute temperature. Warm air is pushed upwards with a pres-
sure difference from the temperature difference:
ptemp ≈ −13hT o (1/To − 1/Ti ) ≈ 13h(Ti − To )/Ti , (14.31)
where To is outside temperature at the top of the stack, Ti is house inside tempera-
ture. At Ti = 20◦ C = 293 K and T = To – Ti ,
ptemp = 0.044hTi . (14.32)
For a two-story house with an attic or chimney at h = 10 m and To = 0◦ C in
winter, the thermal pressure difference is
ptemp = (0.044)(10 m)(−20 K) = −9 Pa. (14.33)
The air infiltration flow rate (m3 /s) from Bernoulli’s theorem is given by
F = area × velocity = A(2p/ρ)1/2 . (14.34)
If stack loss is confined to a total area 0.1 m by 0.1 m, the air loss rate from a
temperature-driven pressure difference of 9 Pa is
F = (0.01 m2 )(2 × 9 Pa/1.3 kg/m3 )1/2 = 0.037 m3 /s. (14.35)
The energy loss rate is
P = ρ F cT/η = (1.3 kg/m3 )(0.037 m3 /s)(1.006 kJ/kg K)(20 K)/1 = 0.97 kW
(14.36)
for electrical heating with η = 1. The fuel rate is 1.2 kW for a furnace with η = 0.8.
This leak costs 24 kWh/day or $2.40/day at 10 c/kWh. For natural gas, it is a loss
of 105 Btu/day = 1 therm/day = $1/day.
14.6.2 Wind-Induced Infiltration

Wind blowing parallel to a surface reduces pressure at the surface due to the
Bernoulli effect. A reduced pressure at the outside surface draws air from the inside
through building leaks. The removed air is replaced by cold winter air or warm
summer air. Large area leaks have laminar flow proportional to the pressure difference
between the inside and outside. Minute area leaks have viscous flow proportional
to the square root of the pressure difference. Since buildings have both laminar and
viscous flow, the flow rate has to be determined empirically with a blower door. A
blower door fan that produces air flow of 1–2 m3 /s can create an overpressure of
100 Pa. A graph of rate F vs. pressure difference p is fit to F = kp n , where n is be-
tween 0.5 for viscous flow and 1.0 for laminar flow and k is a constant proportional
to effective area.1 The curve is extrapolated to a 4 Pa defined pressure difference to
determine the effective loss area.
1
The transition from laminar flow to turbulent flow takes place at higher Reynold’s numbers,
R = vL/η, where v is velocity of air, L is hole diameter, and h is viscosity.
P1: GFZ
Besides measuring leaks, blower doors also discover air leaks by overpressurizing
houses, which forces air to escape through cracks. The trail of smoke coming from a
miner’s smoke stick locates leaks, which are then plugged with polystyrene foam or
calking. Leaks into an attic are discovered with smoke sticks by under pressurizing
a house, that is, sucking air from the attic into the house. In this way blower doors
are used to reduce infiltration losses, which at the extremes of weather, can cost
$1500/year in Maine (heating) and Miami (cooling).
Small blower doors are used to discover leaks in hot air duct systems. The
Lawrence Berkeley group, led by Max Sherman, discovered that a “typical house
with ducts located in the attic or crawlspace wastes approximately 20% to 40% of
heating and cooling energy through duct leaks, and draws approximately 0.5 kW
more electricity during peak cooling periods. Sealing leaks could save close to
1 quad/year nationwide.”2 The LBL group devised a relatively simple repair to fill
the leaks in the ducts. An aerosol sealant is sprayed into a closed, overpressurized
duct system. The excess pressure pushes the sealant to the leaks, where it collects
and plugs them. This is the same approach that is used to seal car radiators, in
which a liquid sealant is added to the coolant, which is pushed into the leaks and
plugs them.
14.6.3 Air-To-Air Heat Exchangers

Superinsulated houses reduce infiltration rates to as low as 0.1 air exchange per
hour, but the resultant air may suffer from indoor pollution and radon. Air-to-
air heat exchangers transfer the heat in warm exiting air to the cold and clean air
entering from outdoors. The temperature efficiency of an air-to-air heat exchanger is
η = Trise /Tin−outside , (14.37)
where Trise is the temperature rise of the incoming air through the heat exchanger
and Tin−outside is the temperature difference between inside and outside air. Tem-
perature efficiency varies between 50–85%, depending on design and weather con-
ditions.
14.6.4 House Doctor Visits

Other measurements and devices provided by house doctors include
r thermoelectric meters to measure heat flow through walls;
r infrared temperature scanners to find and quantify thermal leaks;
r thermocouple temperatures over the 24-h cycle;
r kilowatthours meters on appliances;
2
Raising R values above R3 (English) to R5 and R7 is less effective. The additional savings
from raising R0 to R3, R3 to R5, and R5 to R7 were measured in several cities with these re-
sults: Phoenix (23%/6%/3%), Washington, DC (18%/6%/3%), Minneapolis (17%/4%/3%)
(Sherman and Walker, 1998; Sherman et al., 2000).
P1: GFZ
r solar-flux gain and thermal loss meters on windows;

r lumens/watt measurements on lighting.
14.6.5 Thermal Simulations

Engineers, architects and regulators rely on computer simulations to improve build-
ing designs and determine if a building satisfies state energy regulations. For exam-
ple, results of simulations revealed a $9 million energy savings for a San Francisco
federal building if it were to use natural ventilation without fans or air conditioning.
This would be accomplished by taking advantage of interior heat that is absorbed
during the day on exposed ceiling slabs, which then dissipate heat at night to warm
the building. The DOE-2 and Energy Plus energy loss calculations use heat diffusion
equations that are accurate to 10%. The calculations include the following features:
r solar gain from windows and skylights
r heat conduction through all surfaces (interior and exterior)
r infiltration of air
r heat gain from occupants, lights, and equipment.
An alternative type of calculation uses the analog of electrical RC networks to

determine energy flow and energy storage in buildings. In the analog model voltage
is the analog of temperature, electrical current is the analog of heat flow, capacitance
is the analog of thermal mass, and electrical resistance is the analog of the product of
R-value and surface area (but without i 2 R losses).
14.6.6 Heat Island Mitigation

Urban centers of large cities have seen an increase in average higher summer tem-
peratures prior to the 1950s. The trend has increased air conditioning loads and air
pollution. (See Section 8.7 for heat island models.) The yearly high temperature in
downtown Los Angeles at 4-PM rose 3.5◦ C (7◦ F) in the past 50 years. This is par-
tially driven by the fact that black asphalt reflectivity is as low as 4%, something
that could be mitigated with a shift from green shingles to white shingles on house
roofs, raising reflectivity from 18% to 43% and dropping roof temperature by 10◦ C
(21◦ F). Higher reflection is possible with TiO2 surfaces, which have a reflectivity
of 83%, further reducing roof temperature by another 17◦ C (33◦ F). The LA albedo
could be raised 0.08 by the planting of more trees and an increase of roof reflectivity
over 25% of the city’s area. These actions could lower LA downtown temperature
in 20 years by 2.5◦ C (5◦ F), enough to save 1.5 GWe peak power. In similar fashion,
the United States could reduce the national air conditioning peak load of 100 GWe
by 20% with increased reflectivity.
14.6.7 Building Energy Standards

Many studies show a 25% energy savings on new construction could be realized
for little or no net initial cost, since money could be saved by down-sizing air
P1: GFZ
conditioners and heating plants. About 1.7 million new housing units are built
each year, but this is only 2% of all housing stock and it will take 50 years to take
full advantage of the energy savings. One-half of the states, which account for two-
thirds of new homes, have mandatory energy building codes that are as restrictive
as the standards recommended by the American Society of Heating, Refrigeration
and Air Conditioning Engineers (ASHRAE). California led the nation with its 1975
energy laws, and the state mandated the tightest of US standards (Section 11.4).
California building standards have to be shown to be cost effective for them to
be adopted (Section 16.3). A few cities require existing houses to fulfill a minimum
energy standard before the houses can be resold. A much less restrictive approach
is the use of labels that indicate a house’s energy saving features and failures. New
homes earn an Energy Star if they consume at least 30% less energy than energy code
specifications for heating, cooling, and water heating. This represents a savings of
some $300–$600/year [0.3($1000–$2000/year)] from code level.
14.6.8 Interior Microwave Heat

The ideas presented next are not likely to be implemented, but it is interesting to
consider them. In 1980 Robert Pound suggested heating room occupants directly
with microwaves, similar to warming coffee in a microwave oven. A typical human
radiates about 60 W from a body area of 1.7 m2 . This chemically produced energy
gives us a skin temperature of 34◦ C (93.2◦ F), which is comfortable when 12◦ C
colder in rooms of 21◦ C (70◦ F). It stands to reason that an external 60 W deposited
on human skin would allow a room to be comfortable at 12◦ C lower than standard
room temperature, at a room temperature of 10◦ C (52◦ F). All furnace energy could
be saved for days when outside temperature was at least 10◦ C (52◦ F), and even
lower if one considers free temperature (Section 11.5). The idea is that rooms would
have aluminized wallpaper to scatter microwaves, which would be absorbed by
people and furniture. Microwave ovens operate at 2.45 GHz, but Pound suggested
8 GHz since this frequency penetrates flesh only 10% as deep as oven microwaves.
This approach spreads the 100 W over about 10,000 cm2 for a dose of 10 W/cm2 ,
which is the microwave protection recommended limit in western countries.
14.7 Cogeneration
In 2001, 39% of US energy was used to make electricity and two-thirds of that
amount was ejected into the environment. The 2001 rate for heat waste from
electricity production was
d Q/dt waste = (97.3 quad/year)(0.39 electricity)(0.67 waste) = 25 quads/year,

(14.38)
25% of total US energy consumption. Rejected low-temperature heat is of little

economic use for generating further electricity. However, if properly distributed,
P1: GFZ
14.8. Utility Load Management 365
low-temperature heat could heat cities, as is done in Sweden and Russia. Further-
more, some industries need considerable low-temperature heat in large amounts,
similar to the rates of waste heat from power plants. Dow Corning Corporation
built a cogeneration power plant next to its chemical plant in Michigan plant to
use the rejected low-temperature heat. This conservation measure allows them to
use the rejected heat and sell electricity as a byproduct. Similarly, California uses
cogenerated stream to raise underground temperatures, reducing oil viscosity for
enhanced oil recovery.
Using ideas from Joan and Marc Ross, we compare (1) power and heat from
cogeneration to (2) power from a central station and heat from a boiler. Suppose
a chemical factory needs 1 GWt , of process steam power, which the cogenerator
produces by converting 45% of fuel energy to process steam. In addition, the co-
generator converts 25% of fuel energy to electricity and wastes 30% of fuel heat.
The power yield by the cogenerator for each energy conversion is as follows:
Pcogen−heat = 1 GWt (14.39)
Pfuel = 1 GWt /0.45 = 2.22 GWt (14.40)
Pelec = (2.22 GWt )(0.25) = 0.56 GWe (14.41)
Pwaste = (2.22 GWt )(0.3) = 0.67 GWt . (14.42)
For the case of a separated power station and steam boiler, the power station converts
35% of fuel energy to electricity, and 65% is rejected. To produce the same electrical
power as the cogenerator (Pelec = 0.56 GWe ), the power plant would need a fuel
rate of

Pfuel = 0.56 GWe /0.35 = 1.60 GWt (14.43)
and a waste power rate of

Pwaste = (1.60 GWt )(0.65) = 1.04 GWt . (14.44)
To produce the same 1 GWt , process heat at 80% efficiency, the boiler needs a
fuel rate of

Pfuel = 1 GWt /0.8 = 1.25 GWt . (14.45)
The thermal power for cogeneration is Pfuel = 2.22 GWt , which is less than the sum
of the thermal powers of power plant plus boiler, P’fuel + P’fuel = 1.60 + 1.25 =
2.85 GWt . Hence, cogeneration saves 0.63 GWt , which is 22% of the fuel for a
power plant, plus a boiler. Cogeneration continues to be a viable option, but it will
become less popular now that combined-cycle gas turbines generate electricity at
60% efficiency.
14.8 Utility Load Management

Electrical utility demand is generally two to three times greater between 2 and 5 p.m.
than between 1 and 4 a.m., as seen in Fig. 14.7. Traditionally utilities respond to this
P1: GFZ
Figure 14.7. California 1999 summer peak-day end-use load (GW). The 10 largest building-
sector end uses are shown separately, while the smaller building end users are aggregated
together in remainder of buildings sector. The end-uses are ordered the same vertically in the
graph and the legend. Res = residential buildings, Com’l = commercial buildings. The two
nonbuilding sectors (industry and agriculture/other) are the bottom two segments. Agricul-
ture and other sector includes water pumping, transportation and street lighting (Brown and
Koomey, 2003).
variation by using coal and nuclear base load plants throughout the 24-h day and
natural gas peak-load plants for high demand times. Such decisions are driven by
economics. Base-load plants need longer start-up times, and they have an efficiency
of about 35%. On the other hand, peaking units start up quickly, but they have an
efficiency of only 25%. (Modern combined-cycle gas turbines are 60% efficient. see
Section 10.6.) Base-load plants usually cost more and have to be amortized over the
P1: GFZ
full day, while peak-load plants are less capital intensive and are used only when
needed.
To flatten the demand curve, some power companies have built pumped storage
systems to store base-load energy at night, letting them regain the energy from
hydropower by day with a 20% loss (Section 13.4). Another approach has been
to use time-of-day pricing. In this section we discuss remedies that help flatten the
load curve, reduce the number of power plants (kilowatt savings) and save energy
(kilowatthour savings).
14.8.1 Public Utility Regulatory Policies Act

The 1978 Public Utility Regulatory Policies Act allows small power producers to sell
electricity to electrical utilities or power distribution companies on the basis of
avoided cost. The accounting unit that is employed is the marginal cost of making
electricity, which is the cost to the utility to produce an additional kWh. The util-
ity’s marginal cost, which varies with production rate and time of day, is higher at
higher power levels. It is determined by ranking the utility’s power sources from
the cheapest to the most expensive. If wind or cogeneration power is produced at
5 c/kWh and the utility’s marginal cost of producing electricity is 6 c/kWh, the
small producer can obtain the avoided cost of 6 c/kWh from the utility and make a
1 c/kWh profit. In practice the economics is more complicated because contracts
take into account the attributes of the purchased power (availability, timing, quan-
tity, quality and reliability). The 1978 law opened the gates to the small, decentral-
ized suppliers, such as wind farms and cogenerators. For residential photovoltaics,
the utilities charge a $5/month fee and will cancel power costs, but they will not
pay the integrated cost of a net positive production of kWhs.
14.8.2 Energy Savings

Generation efficiency of base-load plants is higher than it is for peak-load plants.
For every kilowatthour shifted from peak load to base load, the savings in fuel is
calculated as
E fuel = (kWe h)(1/ηpeak − 1/ηbase ). (14.46)
Using ηbase = 60% for the new CCGTs (instead of 32–40% for older plants) and
ηpeak = 25% for peak power plants, the amount of fuel saved is
E fuel = (kWe h)(1/0.25 − 1/0.6) = 2.3 kWfuel h. (14.47)
The fractional savings on fuel for those kWe h shifted from daytime is considerable:
E fuel /E fuel = (E fuel )/(kWe h/ηpeak ) = (2.3 kWt h)/(kWt h/0.25) = 58%.
(14.48)
P1: GFZ
14.8.3 Capital Recovery Fee

Owners of new building usually pay a one-time charge to obtain water service
with a meter. Similarly, industrial customers have to pay such a fee to obtain elec-
tricity with a meter. Utilities use capital recovery fees (CRFs) to help build new
supply-water plants. These fees can be used to reduce wasteful water and electric-
ity practices.
Home owners pay average prices for electricity, based on a utility’s average of
it’s supply costs. This system favors new customers since newer power plants are
usually more expensive. In essence, a price break is given to new customers by old
customers. We explain this outcome by examining the economic reality for older
customers.
Electrical utilities spend $0.50/W for CCGTs and $1–$2/W (or more) for modern
coal and nuclear plants. This amounts to an investment range of $50 to $200 for a
100-W bulb in constant use and perhaps $2000–$8000 for a 4-kW air conditioner
in constant use. If each house were to shift 20% of its peak load of 5–10 kW to the
off-hours, the utility would save 1–2 kW of peak power, saving $500–$2000 per
house.
As an example, assume there are N customers. A 10% increase in customers
is given as N = 0.1N. For the sake of argument, assume new plants produce
electricity at 10 c/kWh, which is twice the cost of 5 c/kWh from old conventional
and hydroelectric plants. (The cost disparity of Cnew = 2C is not true for today’s
CCGTs, which we ignore at first.)
By adding N new customers, the average cost of electricity rises from C to Cavg
according to
Cavg = (CN + Cnew N)/(N + N) = C + (Cnew − C)[N/(N + N)] (14.49)
= 5 c/kWh + (10 − 5)( c/kWh)[0.1N/1.1N] = 5.5 c/kWh. (14.50)
Under this scenario, old-customers’ bills would rise by 10% so new-customers’ bills
would be reduced 50% from 10 c/kWh to 5.5 c/kWh.
The California Public Utility Commission (PUC) does not allow a CRF to encour-
age conservation. Since utilities must add 5–10 kWe additional capacity for each
new house, the CRF could be 50% of the costs incurred when capacity is increased,
or (0.5)(510 kW)($1000/kW) = $2500–$5000 for each new hookup. Customers faced
with this up-front payment would have an incentive to invest in energy-efficient
technologies. If thermal losses are cut in half, air conditioners and furnaces could
be downsized by one-half, saving energy, saving money, and reducing the CRF.
14.8.4 Smart Meters and the Spot Pricing of Electricity

Consumers have essentially no options when buying electricity. A utility charges
some kind of average cost from its many sources of electricity. This approach does
not allow the hidden hand of the market place to match purchases with the large
variety of supplies available. However, utilities offer dual meters to keep track
of consumption during two different time regimes, using different prices during
P1: GFZ
peak and off-peak periods. This is a partial market place measure. These time-of-
use studies showed that a factor of 2 in price encourages the consumer to reduce
consumption during the peak hours by about 10–20%, for a demand elasticity of
−0.15
e d = (d/d)/(p/ p) = = −0.15, (14.51)
1
where demand is d and price is p.
Utilities in UK and Georgia expanded on the dual-meter concept by developing
more complex metering in the 1980s. With the increased availability of inexpensive
computer logic, it is now possible to keep track of electricity consumption as a
function of time with “smart meters” during time intervals of 5 min. An electronic,
digital meter costs about the same as a traditional mechanical meter, about $100,
giving no initial cost at the time of new house construction. The cost of retro-fitting
existing houses can be recovered by negating the more than $10/year cost to read
the meters directly. Smart meters require $10 smart switches, which turn appliances
on and off when responding to several choices for the spot price of electricity. The
smart switches determine the current spot price of electricity, which is set by the
utility. The utilities signal the spot price as a 30-kHz ripple on the 60-Hz line voltage,
or by radio signal. When spot price rises above the price selected by homeowners
on smart switches, any appliances whose switch is in “on position” is turned off.
When the price drops to its former level, the appliances that were turned off will
be turned on, if left in the same position. Smart meters keep track of the use and
price of electricity as it is consumed.
High-power devices with flexible timing (air conditioners, electric water heaters,
dryers) would be excellent candidates for smart metering, and their smart switches
can be over ridden with a flick of the wrist. Many Europeans heat their houses
by using inexpensive nighttime electricity to heat ceramic bricks or an oil bath
in order to give heat by day or night. Refrigerators would need a special switch
that limits down times to a few hours and defrosting to evenings. Adam Smith
would be delighted that smart meters using spot pricing have strengthened the
“hidden hand” that couples cost of need more closely to production. Less beneficial
tasks could be fulfilled with cheaper, off-peak electricity, when time urgency is less
important.
14.8.5 Grid Stability

The August 14, 2003, blackout of eight eastern states and parts of Canada showed
that electrical grids have normal modes that can crash a system. Smart meters
could be a strong force for stabilizing grids and prevent blackouts. When a utility
shuts down one or more large power plants, the utility could quickly raise the spot
price of electricity. Smart switches would quickly respond to the higher prices,
by shedding customers who elect only to pay lower prices. Commercialization of
smart meters is becoming a reality. The 2002 energy act contained a provision to
encourage utilities to read their electrical meters remotely. Utilities could benefit
from smart meters that shift demand from peak to base loads, reducing the need
P1: GFZ
for more power plants. If grid stability was enhanced, the level of reserve power
could be reduced accordingly.
14.8.6 Conservation Voltage Regulation

Voltage regulations in California have produced electricity savings of 3%. In the
1970s, the average California consumer used 118 V, yet the voltage was often much
higher. Prior to 1977, utilities allowed bus-bar voltages to vary between 114 V (peak
power) and 126 V (off-peak power). Regulations by the California PUC reduced
the allowable variation to between 114 and 120 V. We estimate the fractional en-
ergy saved by the PUC voltage control plan. For simplicity, assume a step function
demand profile with peak load the same as base load during a 12-h day. During
12 h of night, only base load plants are in operation. In this scenario, peak-load
plants produce N kWh/day and base-load plants produce 2N kWh/day. Fur-
ther, assume the efficiency of base load plants to be 35% and peak load plants to
be 25%.
Since savings in power (P = V 2 /R) takes place only during high-voltage, off-
peak hours, voltage control does not help the peak load problem. The fractional
power savings in off-peak hours is
P/P = 2V/V = 2(126 V − 120 V)/126 V = 9%. (14.52)
On a daily basis, savings is only 3%, since only 1/3 of electricity is generated
during off-peak hours. The rate of energy consumption before the PUC took action
was
(2N/0.35) + (N/0.25) = 9.7N, (14.53)
and the rate after the PUC acted was
(2N)(1 − 0.048)/0.35 + (N/0.25) = 9.4N, (14.54)
for a fractional energy savings of E/E = (9.7N − 9.4N)/9.7N = 3%. Reduction
in peak voltage not only saves energy but it also lengthens lifetimes of incandes-
cent bulbs by 40%. Electric motor lifetimes were also extended, saved by reducing
overheating at higher voltages.
14.8.7 Leaking Electricity and Standby Power

Field measurements show that 50–100 W leak to ground in typical homes. Using
the lower limit, this corresponds to
(50 W)(100 million homes) = 5 GWe , (14.55)
or 1% of the US average power of 450 GWe . This is the case in other industrial
nations, as well. Instant-on televisions draw 30% of normal power as standby power
when not in use. The computer age has magnified this problem as plugged-in
laptops and printers draw power when not in use. Instant-on audio systems draw
20 W (175 kWh/year) when resting, which is comparable to the average power
P1: GFZ
14.9. Energy Storage 371
of new refrigerators at 60 W (525 kWh/year). An annual sale of 5 million audio

systems requires a growth of 100 MWe /year standby power. The goal is to reduce
standby loss to one W per appliance, with little or no extra cost to manufacturers.
Two-thirds of this loss could be prevented by (a) adding extra switches, (b) moving
switches to the high-voltage side of the line, (c) using capacitor-switches to turn off
equipment, and (d) unplugging equipment when not in use.
14.9 Energy Storage

Over 50% of energy is wasted or rejected. Engineers are not oblivious to energy
waste, but they lack cost-effective measures to make changes. We describe ways to
store energy in batteries, compressed air, flywheels, and magnetic fields. Pumped
hydro storage is discussed in Section 13.4.
14.9.1 Batteries
Traditional lead acid batteries last for 3–5 years and have an efficiency of only 60–
70%. A Sears Diehard 12-V, lead acid battery draws 25 A for about 120 min before
its voltage drops to 10 V. The power and power density of this battery are
P = V I = (11 V)(25 amp) = 280 W (14.56)

P/ma ss = 280 W/24 kg = 12 W/kg. (14.57)
Stored energy and stored energy density are
E = Pt = (0.28 kW)(2.5 h) = 0.7 kWh (14.58)

E/mass = 0.7 kWh/24 kg = 0.03 kWh/kg. (14.59)
A battery produces 280 W (0.4 hp), but the $100 investment stores only 5 c worth
of electricity. The projected energy density for lithium iron sulfide batteries is 0.25
kWh/kg at power densities of 50–200 W/kg, but their operation at 400◦ C causes
complications. Someday advanced batteries might surpass lead acid batteries by a
factor of 15 in energy density and a factor of 5 in power density.
Fuel cells are discussed in Section 15.5 in connection with automobiles, but we
mention them here because of their similarity to batteries. Fuel cells operate on
either hydrogen or methanol. Because hydrogen storage is difficult, fuel cells on
automobiles usually use chemical reformers to prepare hydrogen from methanol (or
gasoline). Ballard fuel cells are impressive in that they develop power densities of
170 to 1000 W/kg, some 50 times that of lead acid batteries. Automotive fuel cells
develop 50–100 kW (70–140 hp) and buses in Chicago, Illinois, and Vancouver,
Canada develop 205 kW (275 hp). Stationary fuel cells in buildings are used as
cogenerators, producing 2 MWe of electricity and using waste heat to warm a
building.
P1: GFZ
14.9.2 Compressed Air

Another way to save energy involves compression of air at night and it’s expansion
by day to drive turbines. A facility in Huntorf, Germany develops 290 MWe over
2 h by pressurizing air to 40 atmospheres in a volume of 0.3 million m3 . The work
needed for an isothermal compression is
p2 p2
dp p2 p2
W= V dp = nRT = nRT ln = p2 V2 ln
p1 p1 p p 1 p1
= (4 million Pa)(0.3 million m3 )(ln40) = 4.4 × 1012 J. (14.60)
At 50% efficiency, the stored energy produces
(0.5)(4.4 × 1012 J)(1 kWh/3.6 million J) = 0.6 million kWh. (14.61)
This agrees with the published value of (290,000 kW)(2 h) = 0.6 M kWh. The
United States is following Germany’s lead by installing a 100 MWe air compression
unit in McIntosh, Alabama, operating at 100 atmospheres.
14.9.3 Flywheels
Modern flywheels have the capacity to spin at 60,000 rpm, with discharge times
ranging from weeks to months. Up to 92% of the energy in a flywheel can be re-
covered, compared to only 60–70% for lead-acid batteries. Graphite fiber flywheels
have an energy density of
(0.5 kWh/kg)(3.6 MJ/kWh) = 2 MJ/kg, (14.62)
which is 50% that of high explosives (4 MJ/kg) and 20 times that of lead acid batteries
(0.1 MJ/kg). Steel alloy flywheels contain only 10% the energy per unit mass of
graphite-fiber flywheels. At first blush, it seems surprising that the maximum en-
ergy per unit mass (which is reached at high speed before rupture) is much smaller
for strong materials, like steel. This phenomenon can be understood by consid-
ering the trade-off between stored energy and tensile strength. The radial, inward
centripetal force needed to contain rotating mass is
R
d F = a dm = (r ω2 )(ρ A dr ) = (ρ Aω2 /2)(R2 − r 2 ), (14.63)
r
where ρ is mass density and A is a small area normal to the radius. For the case of
r = 0, the tensile strength T is exceeded when
T = Fmax /A = ρω2 R2 /2, (14.64)
where R is the maximum radius. The maximum angular frequency at rupture is
ωmax = (2T/ρ R2 )1/2 . (14.65)
The maximum kinetic energy of a cylinder at ωmax is
E KE−max = I ωmax
2
/2 = (mR2 /4)(2T/ρ R2 ) = mT/2ρ = VT/2, (14.66)
P1: GFZ
14.9. Energy Storage 373
where the moment of inertia of the cylinder is I = mR2 /2. Thus, the maximum
kinetic energy is 50% of the tensile strength times its volume. The maximum energy
density is the stored energy divided by its mass (m = ρV):
E KE−max /m = VT/2ρV = T/2ρ. (14.67)
Thus, there is a competition between strength of materials (T) and the necessary
centripetal force to allow mass to rotate at high speeds. If tensile strengths are
similar, the highest kinetic energy per unit mass will be with the lighter material,
which is easier to hold in place (1/ρ). Graphite fibers are stronger in the longitudinal
direction than steel, and they have a density one-third that of steel, a revelation that
led to graphite fiber tennis rackets and sailboat masts.
The theoretical maximum energy per unit mass in a fiber glass wheel is
E KE /m = T/2ρ = (3.5 × 109 Pa)/(2 × 2300 kg/m3 ) = 0.8 MJ/kg = 0.21 kWh/kg,
(14.68)
where silica glass density is 2300 kg/m3 , with a tensile strength of 3.5 × 109 Pa.
Fiberglass energy density is four times greater than that of alloy-steel at 0.046
kWh/kg and eight times higher than a lead acid battery.
Current graphite flywheels carry 0.54 kWh/kg, three times larger than that of
fiberglass, and they are projected to improve to 0.77 kWh/kg. Though expensive,
graphite fiber storage wheels are a good design choice for cars, which are weight-
constrained. However, cheaper fiberglass is used for stationary electrical storage
since weight is not a factor. A fiberglass wheel with a 5-m radius and a 2-m thickness
has been built to contain considerable kinetic energy:
E KE = VT/2 = π R2 HT/2 = π (5 m)2 (2 m)(3.5 × 109 Pa)/2
= 2.7 × 1011 J = 0.08 million kWh. (14.69)
It would take 150 of these flywheels to store 12 million kWh, the amount of
energy produced by 1-GWe over 12 h. The maximum angular frequency at rupture
is
ωmax = (2T/ρ R2 )1/2 = [2(3.5 × 109 Pa)/(2300 kg/m3 )(5 m)2 ]1/2 = 350 rad/s.
(14.70)
This 60-Hz frequency is much less than that of small graphite wheels that rotate at
1000 Hz.
14.9.4 Magnetic Storage

Type I superconductors, such as tin and lead, regain electrical resistance when in
small fields of 0.05 T (500 Gauss). However, type-II superconductors can carry cur-
rents as high as 50 million A/cm2 , producing fields up to 40 T before going normal.
Magnetic fields are confined to the vortices in type-II superconductors until super-
conductivity is quenched. Thus far, the useable currents for high-field, (high-Tc )
superconductors prevent their use for magnetic energy storage, as their amperage
P1: GFZ
is low. Practical considerations dictate a limit of 10–20 T on type-II superconductors.

The energy density u of a 10-T superconductor is
u = μo B 2 /2 = (107 /4π)(J/m3 T2 )(10 T)2 /2 = 40 MJ/m3 = 11 kWh/m3 .

(14.71)
and 44 kWh/m3 at 20 T. Superconductor storage systems are commercially avail-

able to store 3 MJ (0.9 kWh). They are good for smoothing line voltage fluctuations.
A 1–GWe plant operating for 12 h produces 12 million kWh (4.3 × 1012 J). The mag-
netic volume to store this amount may daunt a designer, such a storage system may
be dangerous at 1 million m3 (100 m on a side). We conclude that superconducting
storage is useful for smoothing line fluctuations, but not the case for energy storage.
Problems
14.1 Summer seasonal storage. A house gains energy at 250 W/◦ C, while it is
cooled to 22◦ C from an average temperature of 29◦ C over 6 months. (a) How
many kilowatthour would it take to air condition the house with a COP of
2.5? (b) How much ice, saved from winter, would it take to cool this house?
14.2 Winter seasonal storage. The house above is heated an average of 20◦ C above
outside temperature over a 5-month winter. (a) How much electrical heat
would it take to heat the house? (b) How much summer-warmed water at
30◦ C would be needed?
14.3 Coolth of the night. A shopping center needs 3 MW of air conditioning at the
height of summer with a COP of 3.5 when it is 20◦ C inside and 30◦ C outside.
(a) What is the approximate COP and cooling power needed to night-cool
the shopping center when it is 15◦ C outside? What is the average cooling
load for the day? (b) Ice is made at night with COP = 3. What is the average
cooling load if ice is made overnight and no air conditioners are used? What
is the average cooling load if the cooling system is downsized so that it runs
at the same rate night and day, supplemented with evening ice? What are
the daytime power requirements for these two situations?
14.4 Photonic efficiency. What are the efficiencies of compact fluorescent bulbs
and the sulfur light-guide system?
14.5 Incandescent to CFL. (a) How much power would be saved if incandescent
bulbs using 200 billion kWh were converted to compact fluorescent bulbs? (b)
Is DOE correct in stating that swapping 15-W CFL over 60-W incandescent in
all homes reduces the pollution equivalent of 1 million autos? (c) How much
natural gas is consumed to make 200 billion kWH at 30% and 60% efficiency?
How much CO2 is saved annually by the CFL swap? (d) What is the cost of
a saved kWh for compact fluorescents that are used 2 h/day?
14.6 Plasma frequency in windows. Derive an expression for the plasma fre-
quency. What electron density is needed for total reflection for wavelengths
longer than 1–2 μ?
P1: GFZ
Problems 375
14.7 Low-E windows. Compare energy lost by two windows in Chicago with
6000◦ F days, one with 90% infrared reflectivity and 10% absorption, and the
other with 100% absorption. Do not forget convective losses (Section 11.2).
14.8 Better refrigerators. (a) What is the fractional improvement in Carnot COP
of a refrigerator if its evaporator temperature is lowered from –8◦ C to –12◦ C
and its condenser temperature is lowered from 50◦ C to 38◦ C. (b) What is the
COP if it takes 475 kWh/year to remove 5 GJ/year? What is the maximum
COP possible for a freezer at –7◦ C (20◦ F) and a room at 21◦ C (70◦ F)? What
is its second law efficiency if electricity is obtained with natural gas at 30%
efficiency and it could be obtained at 60%?
14.9 Dual refrigerator-freezer. A refrigerator with a freezing compartment in the
box uses 60 kWh/month, 50% for the refrigerator and 50% for the freezer.
How much energy could be saved if the two tasks were done in separate
boxes, with one cooling system for the refrigerator and one for the freezer?
Assume COP is 20% Carnot based on the freezing compartment tempera-
ture of –7◦ C (20◦ F), the refrigerator temperature of 4◦ C (40◦ F) and the room
temperature of 21◦ C (70◦ F).
14.10 Gas furnace thermodynamics. A gas furnace used for space heating operates
between an ambient temperature of 40◦ F and a task temperature of 110◦ F. For
every therm (105 Btu) of natural gas, the furnace delivers 60,000 Btu. (a) What
is its furnace’s first law efficiency? (b) What is the most efficient system that
could provide the heat, under these temperature conditions? (c) What is the
second law efficiency of the gas furnace?
14.11 Three COPs. What is the COP for home heating with 20% of Carnot for each
of these three cases: (a) a heat pump using 10◦ C ground heat; (b) a heat pump
using 0◦ C lake water; and (c) a heap pump using –18◦ C Alaskan ground.
14.12 Heating bills. How much energy does it take to heat a tight house with
a loss rate of 500 Btu/h ◦ F in the three locations described in problem 4.11
with 3000◦ F day, 7000◦ F day and 10,000◦ F day heating seasons, respectively
(Section 11.3). Consider electrical resistance heating, natural gas furnaces
with η = 80%, and heat pumps.
14.13 House heating. A heating system delivers 5.8 kW of heat on a typical winter
day. One-tenth of the input fuel is lost through incomplete combustion and
passage to the flue to the outside. Of the heat remaining, 20% escapes through
the duct walls and does not reach the living space. (a) What is the conventional
heating efficiency of the furnace? (b) What is the actual heating efficiency of
the furnace? (c) What is the rate of fuel consumption if the fuel is natural
gas? (d) What is the second law efficiency if heat is deposited at 30◦ C and the
outside temperature is 5◦ C?
14.14 LA heat island. If a heat island raises summer temperature by 5◦ C, what
extra amount of energy is needed to cool a house (COP = 2.5) with a loss
rate of 300 W/◦ C over a 5-month summer. What is the total extra amount of
energy needed for the LA air basin with 10 million persons?
14.15 Cogeneration. Compare energy loads for a building using 5 kW of electricity
and needing heat to cover a loss rate of 300 W/◦ C over a heating season of
P1: GFZ
4000◦ C days. Do the calculations to compare: (a) resistance heating with elec-
tricity generated at 35%; (b) heating by natural gas furnace at 80% efficiency;
(c) cogeneration producing electricity at 25% efficiency, and using 70% of the
waste heat.
14.16 Smart meters. A consumer who uses 1000 kWh/month at 10 c/kWh installs
a $100 time-of-day meter. (a) How much of the load must be shifted from the
day at 10–15 c/kWh to the evening at 5 c/kWh to pay for the meter with a
10%/year capital charge? (b) The failure rate is 0.5%/year with a $70 repair
charge. How much load must be shifted to pay for meter servicing?
14.17 Cool wires. A 300-km line at 345,000 V transmits 98% of the power sup-
plied to it. (a) What is the change in the loss rate if the voltage is double?
(b) The cable is cooled underground from 20◦ C to 78 K with liquid nitrogen,
lowering electrical resistance by 4 × 10−3 /◦ C. What is the change in resistance
and power loss for the cooled line? How much power is needed to remove
resistive heat at 20% of Carnot COP?
14.18 Battery and fly wheel storage. (a) How much energy is saved with 80%-
efficient regenerative breaking of a 1 ton car descending a 200-m hill? Same
for a 20-ton bus. (b) What mass battery storage or flywheel (graphite fiber)
is needed to store this energy for each vehicle?
14.19 Aluminum Cans. (a) What fraction of mass and energy would be saved if
aluminum cans were changed from a 2/1 height/diameter ratio to 1/1 to
minimize area? It takes 1 kWh to obtain 15 g of Al for a can. (b) What energy
does it take to obtain an Al atom from bauxite? (c) How much Al, energy and
power would be saved if each US citizen drank 100 cans/year of the new
shape, with 70% recycling?
Bibliography
Brown, R. and J. Koomey (2003). Electricity use in California: Past trends and present usage
patterns, Energy Policy 31, 849–864.
Brown, M., M.D. Levine, J.P. Romm, A. Rosenfeld and J. Koomey (1998). Engineering-
economic studies of energy technologies to reduce greenhouse gas emissions (Five DOE
Laboratory Study), Ann. Rev. Energy Environ. 23, 287–386.
Clark, E. (1986). Cogeneration: Efficient energy source, Ann. Rev. Energy Environ. 11, 275–
294.
de Beer, J., E. Worrell and K. Blok (1998). Future technologies for energy-efficient iron and
steel making, Ann. Rev. Energy Environ. 23, 123–205.
Ford, K., G.I. Rochlin, R.H. Socolow, et al. (1975). The Efficient Use of Energy, American Institute
of Physics Press, New York.
Goldenberg, J., T. Johansson, A. Reddy and R. Williams (1985). End-use global energy strat-
egy, Ann. Rev. Energy Environ. 10, 613–688.
Hafemeister, D., H. Kelly and B. Levi (Eds.) (1985). Energy Sources: Conservation and Renew-
ables, American Institute of Physics Press, New York.
Hirst, E. (1997). Electric utilities in transition, Ann. Rev. Energy Environ. 22, 119–154.
Kreith, F. and R. West (1997). CRC Handbook of Energy Efficiency, CRC Press, Boca Raton, FL.
P1: GFZ
Bibliography 377
McLarnon, F. and E. Cairns (1989). Energy storage. Ann. Rev. Energy Environ. 14, 241–272.
Nadal, S. (2002). Energy end-use and conservation: Appliance and equipment standards,
Ann. Rev. Energy Environ. 27, 159–192.
National Research Council (2001). Energy Research at DOE: Was It Worth It, 1978-2000? Na-
tional Academy Press, Washington, DC.
Office of Technology Assessment (1989). Electric Power Wheeling and Dealing, OTA, Washing-
ton, DC.
———(1993). Industrial Energy Efficiency, OTA, Washington, DC.
Ross, J. and M. Ross (1978). Some energy problem, problems and solutions, Phys. Teacher 16,
272–279.
Rosenfeld, A. (1999). The art of energy efficiency: Protecting the environment with better
technology, Ann. Rev. Energy Environ. 24, 33–82.
Rosenfeld, A., T. Kaarsberg and J. Romm (2002). Technologies to reduce carbon dioxide
emissions, Phys. Today 53(11), 29–34.
Sherman, M. and I. Walker (1998). Can duct tape take the heat?Home Energy 15(4), 14–19.
Sherman, M., I.S. Walker and D.J. Dickerhoff (2000). Stopping Duct Quacks: Longevity of
Residential Duct Sealants, Proc. 2000 ACEEE Summer Study on Energy Efficiency in Buildings,
Washington DC. [LBNL-45423]
Schipper, L., R. Hawarth and H. Geller (1990). US energy use from 1973 to 1987: The impact
of improved efficiency, Ann. Rev. Energy Environ. 15, 455–504.
Schoenung, S., J.M. Eyer, J.J. Iannucci and S. Horgan (1996). Energy storage for a competitive
market, Ann. Rev. Energy Environ. 21, 347–370.
P1: GFZ
15
Transportation
It’s a lean car . . . a long-legged dog of a car

. . . a gray-ghost eagle car.
The feet of it eat the dirt of a road
. . . the wings of it eat the hills.
Danny the driver dreams of it when he sees
. . . . women in red skirts and red sox in his sleep.
It is in Danny’s life and runs in the blood of him,
. . . . a lean gray-ghost car.
[Carl Sandburg, Portrait of a Motorcar]
The romantic rides in Sandburg’s “eagle-car” have changed society. On the one
hand, motor vehicle transportation is an integral thread of society’s fabric. On
the other hand, excess mobility fractures old neighborhoods and families. As our
modes changed from camels to planes to space, our heroes evolved from Marco
Polo to Charles Lindbergh to John Glenn. It took until the beginning of the 19th
century for vehicles to go faster than galloping horses; now they exceed horse
speeds by a factor of 1000. Along with the chaos, this progress mightily increased
the standard of living in developed nations. Transportation now accounts for 28%
of the US energy budget, costing $1200/household for all sectors of society (2001).
The 191 million US drivers spend $400 billion/year on their autos1 , which is 4%
of the GDP and it accounts for 50% of urban air pollution (Section 6.5).2 Global
1
US automobile transportation cost $407 billion in 2001: $167 billion for new and used cars,
$46 billion for repairs, $162 billion for gas and oil, and $32 billion for insurance. [DOT,
2003]
2
Total US NO2 emissions in 2000 were 24.9 million tons (transportation 10.7 Mtons). This
was similar to 1970 data of 26.4 Mtons (transportation 11.4 Mtons). SO2 emissions in 2000
were 18.2 (transportation 1.3 Mtons), much less than 1970 data of 31.2 Mtons (transportation
0.5 Mtons). Lead emissions were considerably reduced from 1970 emissions of 221,000 tons
(transportation 173,000 tons) to 2000 emissions of 4000 tons (transportation 560 tons). Non-
transportation emissions come from industry and electricity generation.
378
P1: GFZ
15. Transportation 379
Figure 15.1. Petroleum consumption by sector. Transportation uses 70% of US petroleum

(Energy Information Administration, 2001).
dependence on petroleum has led the United States into military involvement in
the Middle East.
Energy efficiency of transportation has improved greatly since the oil embargo.
Fuel economy of new cars doubled during the period 1973–2002, from 13.5 to
27.5 miles/gal (mpg) under the Corporate Average Fuel Economy (CAFE) stan-
dards. However, sales of vans, light trucks, and sport utility vehicles now equal car
sales. These two classes of vehicles are 20.7 mpg and 27.5 mpg, giving a total new
light vehicle rate of only 24 mpg. The rate for all light vehicles on the road is smaller
yet, only 19.8 mpg at a time when imported petroleum fills 60% of domestic needs.3
Fuel efficiency of the average pickup truck dropped from its high of 19.2 mpg in
1987 to 16.8 mpg in 2003.
In curbing gasoline consumption, legislative approaches have not worked; an
attempt to raise the standard for all light vehicles to 40 mpg by 2015 failed by 65 to
32 in the US Senate in 2003. With enhanced standards blocked in Congress even
before this vote, the Clinton administration advocated research and development in
its Partnership for a New Generation of Vehicles (PNGV) with a goal of 80 mpg. The
PNGV program was established in 1994 with $1 billion from the government and
a like amount from industry. Progress was made on some PNGV components but
not enough to compete with the Toyota Prius hybrid, which gets 52 mpg in the city
and 45 mpg in the country. In 2002 the Bush-II administration scraped the PNGV
program and established the long-term Freedom Car program based on hydrogen-
powered fuel cells (Section 15.5). See Fig. 15.1 for petroleum consumption by sector
and Table 15.1 for transportation statistics.
3
Average in 2000 of all passenger cars on the road was 22.0 mpg, consuming 532 gallons/car
per year. Light trucks and SUVs on the road averaged 17.5 mpg, consuming 672 gallons/SUV
per year.
P1: GFZ
380 15. Transportation
Table 15.1. Energy transportation statistics.

Vehicle Energy (quads) Percent
autos/motorcycles 9.1 35%
light Trucks/SUVs 6.6 25%
heavy trucks 4.8 19%
buses 0.2 0.7%
air 2.1 8.1%
rail 0.5 1.9%
water 1.3 5.0%
pipelines 0.7 2.6%
other 1 4.0%
Total 26 quads 100%
Transportation consumed 27% of US energy in 2001.

(Department of Transportation, 2004).
15.1 Automobile Energy Basics

15.1.1 Oil Consumption
In this chapter we use SI units for calculations and then switch to English units
recognizing that US consumers drive miles and purchase gallons. At the time of
the 1973 OPEC oil embargo, the United States had 100 million cars and light trucks
with an average fuel economy of 13.5 mpg, traveling 10,000 miles/year. See Fig. 15.2
and Fig. 15.3. In addition, 20 million other vehicles had a smaller fuel economy.
Assuming the total fleet of 120 million vehicles (veh) had an average economy of
12 mpg, total consumption in 1973 was
(1.2 × 108 veh)(104 mi/year)(1 gal/12 mi)(1 bbl/42 gal)

= 2.4 × 109 bbl/year = 6.5 Mbbl/day. (15.1)
Figure 15.2. Average fuel economy of new light–duty vehicles (EIA, 2006).
P1: GFZ
15.1. Automobile Energy Basics 381
Figure 15.3. Automotive fuel economy standards (AFES) and manufacturers’ CAFE levels
(National Research Council, 2002).
The fuel economy for all existing light-vehicles on the road in 2000 was about
20 mpg. That year a greater number of vehicles (210 million) traveled some
12,000 mi/year, consuming
(2.1 × 108 veh)(12, 000 mi/year)(1 gal/20 mi)(1 bbl/42 gal)

= 3.0 × 109 bbl/year = 8.2 Mbbl/day. (15.2)
Total US petroleum consumption is projected to rise from its 2003 level of 20

Mbbl/day to 28 Mbbl/day by 2025, with 65% of the fuel imported. There is little
to hope in the way of decreasing these numbers; the fraction of those persons who
commute by automobile rose from 70% in 1960 to 90% in 2000.
15.1.2 Drag Forces

Thermodynamics and friction constrain internal combustion (IC) engine efficiency
to about 25%. This is further reduced with aerodynamic and rolling drag to 15% at
the wheels. Aerodynamic drag forces Faero are proportional to velocity squared, since
momentum gain (p) of pushed air is proportional to car velocity (v) and it is in-
versely proportional to the time interval (t), which in turn inversely proportional
to velocity. This gives a drag force proportional to v2 :
Faero = ma = p/t α v/(1/v) α v2 . (15.3)
Alternatively, the rate of momentum transfer to a tube of air in front of a car is
Faero = p/t = (m/t)v = (ρ Av)v = ρAv2 , (15.4)

P1: GFZ
where air density ρ = 1.3 kg/m3 and Ais car frontal area. Streamlining cars reduces
effective area with an aero drag coefficient Cd , giving drag force and drag power,
respectively:
Faero = ρCd Av2 /2 (15.5)
Paero = Faero v = ρCd Av /2.3
(15.6)
Some representative Cd values: Volkswagen Rabbit 0.4, Ford Taurus 0.32, Toyota
Prius 0.29, and Ford Probe IV research-vehicle 0.15. Thus, the aerodynamic power
loss of a Taurus with its 2-m2 frontal area, traveling at 30 m/s (108 km/h,
67.5 mph) is
Paero = (1.3 kg/m3 )(0.32)(2 m2 )(30 m/s)3 /2 = 11 kW = 17 hp, (15.7)
(0.746 kW = 1 hp). Because of the oil embargo, the US Congress lowered the
65-mph speed limit to 55 mph in 1974. This now-extinct law targeted only aerody-
namic energy losses, which are relevant at high speeds. If we assume consumption
averaging over all velocities under 65 mph reduces maximum potential savings by
one–half, then gasoline savings are
E/E 65 = (0.5)[1 − (55 mph/65 mph)2 ] = 14%. (15.8)
Rolling resistance drag force and power are as follows:
Froll = Cr mg (15.9)
Proll = Cr mgv, (15.10)
where Cr is the rolling-resistance coefficient, m is auto mass, and g is the acceleration
of gravity. Using Cr = 0.01 for a 1400-kg car, the rolling-resistance power-loss at
30 m/s is
Proll = Cr mgv = (0.01)(1400 kg)(9.8 m/s2 )(30 m/s) = 4.1 kW = 5.5 Hp. (15.11)
Aerodynamic drag predominates at high velocities, while rolling resistance pre-
dominates at low velocities. The two drag forces contribute equally when
Proll = Cr mgv = Pdrag = ρCd Av3 /2, (15.12)
at a transition velocity for our example auto of
vtrans = (2Cr mg/Cd A)1/2 = 21 m/s = 75 km/h = 47 mph. (15.13)
15.1.3 Measurement of Cd and Cr

The aerodynamic drag coefficient can be determined as follows: Measure the time
for a car to decelerate in neutral on a flat road from 60 mph to 55 mph. Since the road
may not be perfectly horizontal, take time measurements in both directions to ob-
tain an average deceleration time. The aerodynamic drag coefficient is determined
from Newton’s second law, ignoring rolling resistance:
Cd = 2Mv/ρ Av2 t, (15.14)
where v2 is the interval average of velocity squared. In our experiment, a Saturn
P1: GFZ
15.1. Automobile Energy Basics 383
slowed from 60 to 55 mph in 8 s. In the calculation, we increased deceleration time

to 12 s to account for rolling resistance in order to obtain the correct value of Cd , 0.3.
The value of rolling resistance Cr is determined by pulling a car at constant velocity
in neutral with a rope and spring-scale. A pull of 30 lb on a 3000 lb (1400 kg) car
gives Cr = 0.01. Engine drag can be measured by leaving the car in gear.
15.1.4 Hill Climbing

The power needed for a 1400-kg car to climb a steep 5◦ hill at 30 m/s is
Pgrav = mgv sin θ = (1400 kg)(9.8 m/s2 )(30 m/s)(0.087)

= 36 kW = 48 hp, (15.15)
which is three times larger than the 11-kW aerodynamic drag at 68 mph. High-
power engines are designed for rapid hill-climbing and rapid acceleration. The total
power delivered at the wheels to drive a Taurus up a 5◦ hill is 11 kW (aerodynamic) +
4 kW (rolling) + 36 kW (hill) = 51 kW = 68 Hp. The total efficiency of an IC engine
and its drive-train is about 15%, giving a gasoline thermal power of 51 kW/0.15 =
340 kW.
Additional power is needed for acceleration. Cars can accelerate at 6 mph/s
(2.5 m/s2 ) attaining a speed of 60 mph in 10 s. This acceleration is harder to maintain
at 30 m/s (68 mph) since it requires
Paccel = Fv = mav = (1400 kg)(2.5 m/s2 )(30 m/s) = 105 kW = 140 Hp. (15.16)
15.1.5 Miles/Gallon or Liters/100 kilometer

When a car travels on level ground at 30 m/s, the aerodynamic drag is 11 kW and
rolling resistance drag is 4 kW for a total of 15 kW. At an efficiency of 15% at the
wheels, this requires burning gasoline at the rate of 15 kW/0.15 = 100 kW. A gallon
of gasoline containing 130 MJ is consumed in a time,
t = E/P = (130 MJ)/(100 kW) = 1300 s = 22 min. (15.17)
During this time, the car travels (30 m/s)(1300 s) = 39 km. The fuel economy of
the car is 39 km/gal = 24 mi/gal. Urban fuel economy is 75% this amount even
though velocities are lower, since because of stopping and starting fuel economy
is lowered. Outside the United States the inverse mileage unit of liters/100 km
is used, which is similar to efficiency. It takes 340 s to consume a liter of gasoline
providing 35 MJ while traveling (340 s)(30 m/s) = 10 km. This gives a fuel economy
of 1 l/10 km = 10 l/100 km.
To the first order, fuel economy is essentially inversely proportional to auto mass,
which is not surprising since energy losses from stopping, hill climbing and rolling
resistance are proportional to mass. Aerodynamic drag is almost proportional to
mass since frontal drag area of a cubic car is proportional to m2/3 . A main goal
of CAFE regulations (Section 15.2) has been to improve fuel economy with re-
duced mass (and improved technology) while maintaining vehicle safety and per-
formance.
P1: GFZ
15.1.6 Trends
Beyond CAFE, the present major thrust for improved fuel economy cars is driven
more by environmental concerns than by national security threats on oil imports.
For example, California mandated a fraction of its auto market be filled by cars with
very low or zero fuel emissions.4 The average power of IC-engines in the United
States dropped from 140 horsepower (hp) in 1975 to 100 hp in 1980, but then rose
back to 140 hp by 1991. From 1991 to 2003, engine power in medium-sized cars
rose from 142 hp to 173 hp, and in light trucks it rose from 158 to 197 hp. Engines
were downsized from 4900 cm3 (300 in3 ) in 1975 to 3000 cm3 (180 in3 ) in 1991 as
power density rose 65% from 0.5 hp/in3 in 1975 to 0.82 hp/in3 in 1991. The average
mass of new US cars dropped from 4050 lb in 1975 to 3100 lb in 1985, but then rose
20% by 2000. The trend toward bigness and prowess is on the rise as the weight
of pickups grew from 4000 lb in 1990 to 4600 lb in 2003, while acceleration time to
attain 60 mph dropped from 12 s to 10 s.
Improvements under CAFE came from the following measures:
r mass downsizing of 25%
r electronic engine controls for more efficient combustion
r five-speed manual transmissions
r fuel-injection without a carburetor
r four values per cylinder
r front-wheel drive, reducing drive-train losses
r improved aerodynamics, lowering Cd from 0.4 to 0.3.
Since IC engines can be only marginally improved, there might, at some point, be
a departure from complete dependence on IC engines. Such an idea was considered
heresy a decade ago. A more likely shift is one to hybrid cars that get 50 mpg in
cities (Section 15.4) and, perhaps some day, the hydrogen-powered fuel cell cars
(Section 15.5). For now, options envisioned by the PNGV include very light cars
made with carbon fiber, small diesel engines, compressed natural gas engines,
ethanol/methanol engines, the electric plug-in car, and energy storage in advanced
batteries and flywheels. Super-cars could get 80 mpg with vastly reduced emissions,
but would such cars actually be purchased?
15.2 Corporate Average Fuel Economy (CAFE)

CAFE mandates that manufacturers comply with a fleet-average fuel economy of 27.5
mpg. Since the gallon parameter is in the denominator, fleet-average fuel economy
is not a simple average of individual fuel economies. Consider the average fuel
economy of a 20-mpg car and a 10-mpg car. If both cars traveled 20 miles, the total
4
August 12, 2003, General Motors and Daimler-Chrysler dropped their law suit against
California, accepting the concept of low and zero emission standards. The level of imple-
mentation is in a state of flux.
P1: GFZ
15.2. Corporate Average Fuel Economy (CAFE) 385
amount of gasoline consumed would be 1 + 2 = 3 gal, for an average of 40 mi/3 gal,

or 13.3 mpg, not (20 + 10)/2 = 15 mpg. Since the guzzler’s mpg dominates the fleet-
average fuel economy, manufacturers are encouraged to improve guzzlers more
than already efficient cars. The average fuel economy for our two cars is obtained by
averaging the inverse of fuel economy:
1/fuel economy = (1/10 + 1/20)( gal/mi)/2 = 0.075 gal/m, (15.18)
with an average fuel economy
fuel economy = 1/0.075 gal/mi = 13.3 mpg. (15.19)
The inverse average fuel economy for a fleet of cars is

−1
Ffleet = n /Fi ,
i i
(15.20)
where ni is the fraction of cars in the ith subclass with fuel economy Fi .
In 2000, a panel of the National Research Council estimated that increased fuel
economy under CAFE saves the US 2.8 Mbbl/day. This estimate was not obtained
by merely doubling fuel economy, because light trucks and SUVs consume at a
rate midway between the 1973 13.5 mpg and the CAFE 27.5 mpg. Nonetheless, we
ignore the SUV effect to examine a larger point. The first doubling of fuel economy
cuts gasoline consumption in half. Unfortunately, a point of diminishing returns
undercuts further doublings. Assume national gasoline consumption is
G = C/F (15.21)
where C is the total miles traveled by U.S. cars, which is a constant, and F is
the fleet-average fuel economy. Doubling the fuel economy to CAFE’s 28 mpg
reduces consumption to C/2F , saving C/2F . A second doubling to 56 mpg reduces
consumption to C/4F , saving an additional C/4F . A third doubling to 112 mpg
reduces consumption to C/8F , saving an additional C/8F . With each doubling,
the effect on fuel economy (for example, 56 mpg) is to make it twice the previous
fuel economy (28 mpg), while savings are half as much (C/4F ) as the previous
savings (C/2F ). If gas consumption G = C/F is 8 Mbbl/day with today’s fleet,
then doubling fuel economy to 56 mpg would save 4 Mbbl/day. A second doubling
to 112 mpg would save 2 Mbbl/day, and the third doubling to 256 mpg would save
1 Mbbl/day, clearly a diminished return.
15.2.1 Gas Guzzlers

We examine the improvement of two cars, one at 10 mpg and the other at 20 mpg.
If the 20-mpg car alone is improved to 21 mpg, the fleet average increases by 0.22
mpg to 13.5 mpg. On the other hand, if only the 10-mpg car is improved to 11 mpg,
the fleet average increases by 0.86 mpg to 14.2 mpg, four times the improvement of
the former case (0.86/0.22 = 4). The factor of 4 is obtained by taking the differential
of the inverse fuel economy, giving the change in the inverse fuel economy for one type
P1: GFZ
of car,
(1/F ) = −F /F 2 . (15.22)
The ratio of energy savings of F = 1 mpg for two types of cars (guzzler and saver)
is
guzzler/saver = (1/102 )/(1/202 ) = 4. (15.23)
To discourage purchase of inefficient autos, the 2000 Gas Guzzler Tax triggers a
$1000 guzzler tax on a 22-mpg car (but not on SUVs) and $7700 on a 12.5-mpg car.
15.2.2 Feebates
Because standards and taxes are unpopular, an alternative approach to curbing
fuel consumption was suggested; penalties on guzzlers and encouragement for
savers. A balance point of 28 mpg was proposed by Jonathan Koomey and Art
Rosenfeld (Lawrence Berkeley National Laboratory). They suggested rebates of
$970 for Ford Escorts (35 mpg) and $1250 for Honda Civics (37 mpg), and a $4750
penalty for a Ferrari (15 mpg). On the basis of 1987 sales, $3.4 billion would be paid
in fees and $1.7 billion would be disbursed as rebates. This feebate scheme was not
revenue-neutral (revenues = benefits) to the government, but the structure could
be so modified.
To put these rebates and penalties into perspective, we estimate the fuel cost to
run $2.50/gal Civic and a Ferrari over a 150,000-mile lifetime:
Civic : (150,000 mi/37 mpg)(2.50/gal) = $10,100 (15.24)
Ferrari : (150,000 mi/15 mpg)(2.50/gal) = $25,000. (15.25)
Future gasoline payments should be discounted to the present since we can invest
money in the present to spend later (Section 16.2). The present net cost energy cost
for the Ferrari is sum of the present value of gasoline (about $10,000) and feebate
penalty ($4850), for a total of about $15,000 (at the time of purchase). The Civic’s
net energy cost is much smaller at $2750 ($4000 (gasoline) – $1250 (feebate)). A
major difficulty with feebates is that they penalize large US cars and rebate small
Japanese cars. The fuel cost of a 50-mpg Toyota Prius hybrid would be $7500.
15.3 IC Engines
In the IC engine cycle there is little time for heat transfer during compression and
expansion strokes since they take place in 5 ms. For this reason compression and
expansion strokes are essentially adiabatic, since very little heat transfer takes place
during the very quick stroke. Adiabatic strokes have no heat transfer, Q = 0 with
pVγ = constant (with γ = 1.4). The combustion cycle takes place so quickly that it
is essentially a constant volume process. Similarly, the exhaust cycle is a constant
volume process since the opening of an exhaust valve quickly reduces pressure
P1: GFZ
15.3. IC Engines 387
T3
QH
4
Qc
T1
1
V
Figure 15.4. The auto’s Otto cycle. The air fuel mixture is compressed adiabatically (1–2),
the explosion takes place quickly at constant volume (2–3), the adiabatic expansion stroke
performs work (3–4), and the exhaust valve releases pressure quickly (4–1). The dashed lines
represent constant temperature processes
to one atmosphere. Thus, the Otto cycle consists of two adiabatics (compression
and expansion) and two constant-volume processes (combustion and exhaust). In
practice the sharp corners on the pV plane are rounded (Fig. 15.4). The Toyoto Prius
uses the Atkinson cycle improvement, which has an expansion ratio that is higher
than the compression ratio, allowing more energy to be obtained from the power
stroke.
The efficiency of a heat engine is given as
η = W/Qin = (Qhot − Qcold )/Qhot = 1 − Qcold /Qhot . (15.26)
For the Otto cycle, transferred heat is Q = ncv T where n is the number of
moles, c v is the constant volume specific heat, and T is the change in temperature.
Combining these facts in problem 15.6, gives the Otto cycle efficiency,
η = 1 − r 1−γ = 1 − r −0.4 , (15.27)
where r is volume compression ratio = Vinitial /Vfinal . Large compression ratios
raise temperatures higher, increasing efficiency. Compression ratios of 8 and 10
give η = 56% and 60%, respectively. In practice the efficiency of IC engines when
unattached to the rest of drive train is about 25%, or about one-half the theoretical
value. The efficiency at the wheels is closer to 10–15%, as IC engines are usually
over powered for much of what they do. (This problem is solved with hybrid
cars in Section 15.4.) Very high compression ratios should be avoided because high
P1: GFZ
temperatures create NOx , which combines with unburned hydrocarbons and sun-
light to create smog. On the other hand, small combustion ratios fail to completely
burn the fuel, releasing unburned hydrocarbons. Burning gasoline creates 20%
more molecules than the starting mixture, giving an extra push in the expansion
cycle,
C8 H18 + 12.5 O2 ⇒ 8 CO2 + 9 H2 O. (15.28)
Diesel engines in trucks have compression ratios of 14–20, raising temperature
enough to initiate explosions without spark plugs. The constant volume combus-
tion process of the Otto cycle is replaced with the constant pressure burn of the
Diesel cycle since less combustible diesel fuel burns more slowly. The Otto cycle
IC engine reigns supreme in spite of diesel engines, Brayton cycle for gas-turbines,
Wankel rotary engines and external combustion Stirling engines.
15.4 Hybrid Cars

15.4.1 Hybrid Cars
Hybrids double fuel economy of urban driving for these reasons: (1) Traditional IC
engines are designed for high acceleration and high speeds, forcing most driving to
be done at nonoptimal conditions. By operating two power plants (IC and electric)
in parallel, hybrid cars can have smaller IC engines. (2) Electric motors are more efficient
than IC engines, a result that is useful in less energy-intensive urban traffic. (3)
Regenerative breaking saves energy of motion to the battery. (4) An IC engine periodically
recharges its battery for use in urban traffic.
The 2003 Prius has an electric motor (supplying 33 kW, 44 hp) and an IC engine
(70 hp, 1500 cc, 4 cylinder, 16 valves) to obtain 52 mpg in the city and 45 mpg on the
highway. The nickel metal-hydride, 274-V battery is guaranteed for 8 years, and
its volume is only 15% of that required for the all-electric car. The smaller, 2-seat
Honda Insight obtains 61 mpg in the city and 68 mpg on the highway. Ford Motor
Company now produces a hybrid SUV that gets 35 mpg in the city.
15.4.2 Regenerative Breaking

An electric motor can be used as a generator to recover energy. By spinning a coil
of wire in a magnetic field, current is induced to charge a battery. This concept is
called regenerative breaking and it has long been used by buses in Rio de Janeiro to
save energy of motion and potential energy from descending hills. In 2000, New
York City purchased 125 diesel-electric hybrid buses with regenerative breaking.
Since New York City is relatively flat, the main energy saving is from saving kinetic
energy. About 50% of kinetic energy is available for recycle electricity from regen-
erative breaking, but perhaps 75% could be available through a mechanical system.
Stopping a car at 25 mph (11.1 m/s) could save motion energy in the amount
E regenerate = (0.5)(mv2 /2) = (0.5)(0.5)(1400 kg)(11.1 m/s)2 = 0.043 MJ. (15.29)
P1: GFZ
15.4. Hybrid Cars 389
Stopping a car on the highway at 65 mph can save 0.3 MJ. We compare these
savings with gasoline energy. A gallon of gasoline delivers 130 MJ, but an IC engine
at 10% efficiency in the city delivers only 13 MJ/gal to the wheels. To save 1 gal
of gasoline a car must stop 300 times in city traffic (13 MJ/0.043 MJ per stop) and
43 times on the highway (13 MJ/0.3 MJ per stop). For example, if a car stopped
twice in a mile in urban traffic, a trip of 150 miles would save 1 gal. In urban traffic
a CAFE car gets 20 mpg, burning 7.5 gal in 150 miles. Thus, regenerative breaking
saves 1 gal in 7.5 gal for a 13% savings.
Regenerative breaking saves a similar amount for cars traveling down small
mountains. When a car goes down a 1000 foot (300 m) hill, regenerative breaking
saves
ηmgh = (0.5)(1400 kg)(9.8 m/s2 )(300 m) = 2.1 MJ. (15.30)
This commuter burns 1 gal/day on a 30-mile round-trip for a savings of (2.1
MJ)/(20 MJ per effective gallon) = 10%. Regenerative breaking is most attractive
for buses because of their large mass.
15.4.3 Electricity versus Gasoline

Regenerative breaking and electric cars might force a shift in energy units from
miles/gallon or liters/100 km to kWh/mile. If a car loses 15 kW from aerodynamic
and rolling drag at 30 m/s (68 mph), a trip of 1 km would consume electrical energy
at the rate of
E elec = Pt = (15 kW)(1000 m)(1 s/30 m) = 0.5 MJ/km = 0.14 kWh/km
= 0.22 kWh/mi. (15.31)
If we consider the cost of fuel only, electricity is cheaper than gasoline. At
10 c/kWh, it costs 2.4 c/mile for electrical energy, while gasoline costs 10 c/mile
(25 mpg at $2.50/gal). The 200 million US vehicles, traveling 12,000 mi/year, each
require
(200 million)(12,000 mi/year)(0.22 kWh/mi) = 5.3 × 1011 kWh/year. (15.32)
This amount is increased to allow for energy losses, making total electrical need
at least 8 × 1011 kWh/year. Since a 1-GWe plant produces about 7 × 109 kWh/year,
it would take 115 1-GWe power plants, 20% of the US grid, to sustain an all-electric
US vehicle fleet.
A 30-mpg gasoline car consumes energy at a rate
E distance = (130 MJ/gal)(1 gal/30 mi) = 4.3 MJ/mi = 2.7 MJ/km. (15.33)
This gasoline-powered car consumes 5.4 times the energy of the electric car at 0.5
MJ/km. The electric car did better than the gasoline car because electrical motors
are 90% efficient as compared to 15% for cars. However, this comparison is one of
gasoline energy to stored electrical energy. If the efficiency of a power plant is 33%,
the electric car advantage drops from 5.4 to 1.8. If a combined cycle gas makes
P1: GFZ
electricity at 60% efficiency, the electric car’s advantage rises upward to 3.2. The
favorable efficiency of electric cars would be decisive except for the inefficiencies,
lifespan and cost of battery storage. This is why hybrid and fuel cell engines have
rejuvenated the prospect of electric car use. The ability to generate electricity on
board greatly reduces battery requirements.
15.4.4 Flywheel Storage

Portable flywheels (Section 14.12) have been used to store the energy gained
through regenerative breaking in flywheels with 20-cm diameter and 30-cm thick-
ness. These flywheels were spun at 60,000 rpm with 92% recycle efficiency. A power
density of 8 kW/kg was obtained, which is several times that of IC engines and 100
times that of batteries. Using flywheel energy densities of 0.02–0.04 MJ/kg, a 10-kg
flywheel is sufficient for regenerative breaking at highway speeds and a 100-kg
flywheel is sufficient for 300-m descents. Flywheel batteries are attractive but their
development is not yet mature.
15.4.5 Cost of Conserved Energy

How much extra money would we be willing to spend for a 50-mpg auto-
mobile as compared to a 27.5-mpg CAFE car? The 27.5-mpg car that travels
15,000 mi/year consumes 545 gal/year, while a 50-mpg car consumes 300 gal/year,
saving 245 gal/year. If the extra investment for a Prius is $4000, the cost of the extra
loan and repayment in constant dollars (without inflation) is about $500/year for
10 years (Section 16.3). This puts the cost of conserved energy for the 50-mpg car at
annual cost/annual energy saving = ($500/year)/(245 gal/year) = $2/gal.
(15.34)
Since the price of gasoline is $2.50 to $3.50 per gallon in 2006, we conclude that
purchasing the 50-mpg car is favorable. If the car lasts two decades and is driven
20,000 mi/year, it is a very good purchase. For those living in Europe and Japan
and pay $5 or more per gallon, the investment is clearly worthwhile.
15.5 Hydrogen Fuel-Cell Car

15.5.1 Fuel Cells
Air pollution from automobiles is a major problem (Section 6.5), but the fuel cell
car might someday dramatically improve matters. All-electric cars have not suc-
ceeded because of their need for expensive, heavy batteries. The traditional lead
acid battery has limited capacity to store energy, which in turn limits its travel
range. In addition, a lead acid battery has a relatively short replacement times of
3–4 years, while requiring recharging times of several hours. A likely alternative
is the fuel-cell car, which has been used on a trial basis with urban buses. Because
P1: GFZ
15.5. Hydrogen Fuel-Cell Car 391
hydrogen storage is difficult and not readily available, today’s fuel cell cars and buses
prepare hydrogen on board from methanol or gasoline in reformers. Fuel cells convert
hydrogen into electricity in a battery-like device of the type used for space flights.
Large cost reductions have made fuel cell electric generators an economic pos-
sibility, but, according to the National Academy of Sciences (2001), “commercial
application of fuel cells for passenger vehicles is at least 10–15 years in the future.”
All hydrogen sources have drawbacks (fossil fuels, solar, nuclear, alternative fuels).
Perhaps the easiest approach is to improve the present fleet of automobiles. The
2002 National Academy study on CAFE projects that fuel economy improvements
12–42% above the CAFE standard can be cost effective.
Fuel cells convert fuel into electrical energy at “cold combustion” temperatures
of 80◦ C by combining hydrogen and oxygen to make water. Fuel cells use platinum-
based catalysts that strip electrons from hydrogen and transfer them to an electrical
circuit, while residual protons penetrate a membrane, going into an electrolyte to
await another electron. After the electron has done its work, it returns in the closed
circuit to combine with the proton to make hydrogen, which combines with oxygen
in the electrolyte to form water. The proton-exchange membranes are polymers,
similar to Teflon (Nafion in a Gore-Tex membrane) that efficiently conduct protons,
but are impermeable to water, hydrogen, and oxygen. Many one-volt stages are
combined in series to develop larger voltages.
Fuel cells are much lighter than batteries. Ballard fuel cells produce 32 kW in a
45-kg package, for a specific power of 700 W/kg, 40 times that of lead acid batteries.
Automotive units produce 50–100 kW (70–130 hp), while bus units produce 205 kW
(275 hp). The efficiency of a fuel cell is about 50%, obtaining the equivalent of 80
mpg. Fuel cells operate efficiently at partial load in urban traffic, which is a problem
for IC engines. Fuel cells that are run on hydrogen are emission-free, but because of
problems with gasoline chemistry, methanol is more likely to be used in fuel cells.
Methanol fuel cell efficiency drops to 27% and emits CO2 , which California counts
as half credit in its zero-emission standard.
Fuel cells have positive features, but what are their negatives? To create a nation
of fuel cell cars would be difficult. The prime impediment is the extra cost, perhaps
$10,000 per fuel-cell car. But as economic improvements have been dramatic, the
prospects of fuel-cell cars have improved. Hydrogen fuel cells in the space shuttle
first cost $500,000/kWe , but this dropped to $200/kWe for two reasons: The plat-
inum content was reduced by 90–95% and proton-exchange membranes that allow
cooler operation were invented.
15.5.2 Hydrogen
Compared to copious pollutants and CO2 from fossil fuels, hydrogen’s only
byproduct is water, although high-temperature flames create NOx . A hydrogen
economy has many boosters, but it also has detractors. Calculations in 2003 showed
that hydrogen could cool the stratosphere, enhancing the chemistry that destroys
ozone. Since hydrogen gas and its flame are odorless and colorless, there are ques-
tions about the safety of a hydrogen economy. A hydrogen car economy is uncertain
P1: GFZ
because onboard fuel storage is difficult. On the plus side, hydrogen can be trans-
ported by pipeline. Its energy volume density is 30% that of natural gas (325 vs.
1000 Btu/ft3 ) but its mass is only 12% that of natural gas, giving it 2.7 times the en-
ergy per unit mass. Storage of hydrogen in is being considered in metal hydrides, in
high-pressure tanks (104 psi) and in carbon nanotubes. All of these approaches add
cost and mass, reducing hydrogen’s net energy density. Hydrogen can be produced
in a variety of ways, as we now describe.
15.5.3 Coal Gas

Old-time coal gas contained hydrogen, methane, and carbon monoxide obtained
from the reactions
3C + 3H2 O ⇒ CO2 + CO + H2 + CH4 (15.35)
C + H2 O ⇒ CO + H2 . (15.36)
15.5.4 Steam Reforming

Today’s hydrogen is mostly produced from natural gas (methane) by steam re-
forming at $5/MBtu
2H2 O + CH4 ⇒ 4H2 + CO2 . (15.37)
A total shift to hydrogen through steam-reforming could be difficult since it
would require the United States to double its natural gas production rate from
20 TCF/year to 40 TCF/year. To avoid CO2 release, hydrogen would have to be
produced with solar, wind, or nuclear energy using the cycles described below.
15.5.5 Thermolysis
The thermal splitting of water with high temperatures over 2500◦ C is not effective
since the accompanying reverse reaction is rapid. It would be more effective if it
were possible to directly split water with solar photons over 1.23 eV, but such a
method is difficult. Progress has been made through the use of a water bath on top
of a semiconductor surface to absorb sunlight and directly produce hydrogen.
15.5.6 Thermochemical
Solar concentrators have been used to thermally split water using a sulfur iodine
cycle, a sulfate cycle, a bromide cycle, and a chromium–chlorine cycle. This last
cycle is shown below:
50◦ C 2CrCl2 + 2HCl ⇒ 2CrCl3 + H2 (15.38)
◦
900 C 2CrCl3 ⇒ 2CrCl2 + Cl2 (15.39)
◦
750 C 2H2 O + 2Cl2 ⇒ 4HCl + O2 (15.40)
P1: GFZ
15.5. Hydrogen Fuel-Cell Car 393
Another cycle, one using sodium borohydride in a water solution, makes hydro-
gen on board without hydrogen storage. In an experiment this approach propelled
a fuel-cycle car 300 miles on one tank, but at a higher cost.
15.5.7 Electrolysis
Hydrogen electrolysis costs $30/MBtu from conventional electricity and $50/MBtu
from photovoltaic (PV) electricity, which is 6–10 times the cost of hydrogen from
steam-reformed natural gas. The thermodynamic potential for electrolysis of water
is 1.23 V at 23◦ C. Electrolysis operates at 1.8 V because of voltage losses and higher
operating rates, giving an efficiency of 1.23 V/1.8 V = 70%. The energy available
in a kilogram of hydrogen is
(6.0 × 1026 H atoms/kg)(1.23 eV/H atom) = 7.4 × 1026 eV/kg = 120 MJ/kg
(15.41)
The energy of 1 kg of hydrogen is about that of a gallon of gasoline (130 MJ/gal),

which is 3.5 times heavier than 1 kg of H. The amount of electricity to make 1 kg
of H at 70% efficiency is
(120 MJ/kg H)/(0.7 × 3.6 MJ/kWh) = 48 kWh/kg H. (15.42)
The cost of industrial electricity at 5 c/kWh gives a cost of $2.40/kg H, equivalent

to $2.40/gal gasoline. Thus, H electrolysis might seem competitive with gasoline,
but we must add the cost of the electrolysis plant and its operation, hydrogen
storage and the higher cost of a fuel cell generator. Steam-reforming natural gas is
cheaper at $0.40/kg H, but the process relies on fossil fuels. The efficiency of elec-
trolysis is enhanced at higher temperatures. The Bush administration is considering
building a nuclear reactor to supply both the electricity and the higher temperatures
to do this.
15.5.8 Silicon-Photovoltaic Hydrogen Autos

Photovoltaic (PV) electricity costs about 25 c/kWh (Section 13.2) under a cloudless
sky at a cost of $5/peak watt. At this rate the hydrogen equivalent of gasoline
is $11/gal versus the present cost of $2. However, hydrogen fuel cells with an
efficiency of 50% are three times more efficient than IC engines. This factor of 3
effectively reduces solar hydrogen fuel to under $4/gal. This implies that PV costs
must be reduced by a factor of 2–3 to be competitive with gasoline fuel, but the cost
of fuel cells and hydrogen distribution must be included in the adjustment. This
result may preclude silicon single-crystal PVs for electricity, but thin-film silicon or
amorphous silicon might be competitive.
How large a PV area would be necessary to be able to supply the nation’s cars
and small trucks? To answer this requires a series of calculations: A fuel cell car
running 12,000 mi/year on hydrogen at 80 mpg (equivalent) consumes a gasoline
P1: GFZ
equivalent of
(12,000 mi/year)(1 gal/80 mi) = 150 gal/year, (15.43)
with an energy (150 gal/year)(130 MJ/gal) = 20 GJ/year. A national hydrogen
auto economy in 2010 would need
(250 million vehicles)(20 GJ) = 5 × 109 GJ/year = 5 EJ/year. (15.44)
Peak solar flux s1 in the southwestern United States at noon is 1.0 kW/m2 in
the summer and 0.6 kW/m2 in the winter (Section 12.1). Using 14 h of daylight
(T = 28 h) in summer and 10 h in winter (T = 20 h), the daily integrated solar
fluxes are
Isum = s1 T/π = (1 kW/m2 )(28 h)/π = 8.9 kWh/m2 day (15.45)
2 2
Iwin = s1 T/π = (0.6 kW/m )(20 h)/π = 3.8 kWh/m day. (15.46)
Using the average of summer and winter fluxes, the number of kWh of solar energy
available over the year is
(6.4 kW/m2 )(365 day/year) = 2300 kWh/m2 year, (15.47)
in agreement with measured data. Single-crystal PV net efficiency of 10–15% and
electrolysis with an efficiency of 70% gives hydrogen energy from PV electrolysis
of
(2300 kWh/m2 year)(3.6 MJ/kWh)(0.15)(0.70) = 0.87 GJ/m2 year. (15.48)
Hence, the single-crystal PV hydrogen auto economy requires a PV area of
(5 × 109 GJ)(m2 /0.87 GJ) = 5700 km2 . (15.49)
Land needs of infrastructure (roads, electrolysis, and so on) could double this to
12,000 km2 . This area would be doubled or tripled with amorphous or thin-film PVs.
15.6 Safety
In 2002 there were 42,815 fatalities on the highways, compared to 609 by air and
951 by trains (DOT, 2004). The number of injuries on the highways was 67 times
higher than the fatalities at 2,892,057 in 2002. 40% of the highway deaths were related
to alcohol abuse. The number of traffic deaths per mile for light vehicles was 1.3 per
100 million vehicle miles (4 × 104 deaths/3 × 1012 vehicle miles). This is dramat-
ically lower than the 1935 figure of 16 per 100 million vehicle miles. However, as
total number of vehicle miles/year continues to rise, the total number of deaths
remains relatively constant (40,000–50,000 deaths/year between 1982 and 2001).
The number of deaths due to alcohol dropped from 25,000 in 1982 to 17,000 in 1995
as a result of increased public awareness. As the mass of US cars dropped 20% from
4300 to 3400 pounds between 1975 and 2000, the mortality rate/mile dropped 45%,
indicating that smaller cars are not necessarily less safe than larger cars (Fig. 15.5).
P1: GFZ
15.6. Safety 395
Figure 15.5. Light vehicle fatality rate (small dots) and fuel economy, 1966–2000. Fatalities
per 100 million vehicle miles and miles per gallon (National Research Council, 2002).
The number of autorelated fatalities during an 80-year lifetime is

(40,000/year)(80 year) = 3 million, (15.50)
or 1% of the US population of 290 million. This represents a surprising high risk
factor that is generally ignored, and it does not include the risk of surviving an auto
accident, that is, of being among the walking wounded. The average probability of
dying in one of the US’s 15 million accidents/year is
(40,000 accident deaths)/(15 million accidents) = 0.3%. (15.51)
Automobiles are now designed with more crush space, a design feature that
extends the duration of an accident. This reduces deceleration, which mitigates
the severity of accidents. With crush time extended to 0.1 s, deceleration of cars
traveling at 30 m/s (68 mph) is reduced to 30 g’s. This reduction can save lives.
Frontal air bags enhance safety, and they can be extended to side air bags for $350–
$650.5 Another approach to safer design is antilock brakes, which take advantage
of the 35% larger coefficient of static friction for rolling as compared to the smaller
coefficient of kinetic friction for sliding. Anti-lock systems maximize braking ability
by releasing the brakes just before skidding. This allows the car to continue rolling.
The stopping distance on dry concrete for rolling with static friction (μs = 1.2) at
30 m/s (68 mph) is
x = v2 /2μs g = (30 m/s)2 /(2)(1.2)(9.8 m/s2 ) = 38 m = 125 ft. (15.52)

The stopping distance is 57 m (190 ft) for skidding with kinetic friction coefficient
μk = 0.8. For wet pavements, the rolling stop distance is 57 m (μs = 0.8) and the
5
Side air bags reduced deaths from side collisions by 37%, from 248/105 accidents to 157/105
accidents.
P1: GFZ
skidding stop distance is 77 m (μk = 0.6). However, the NAS concluded, “There is
no evidence that anti-lock brakes have affected overall crash rates.”
15.6.1 Light Versus Heavy

Large pickups and SUVs inflict up to four times the number of fatalities per ac-
cident on small car occupants as small cars cause to passengers in large cars in
accidents. The advent of General Motors’s 6400-lb Hummer (downsized to 4700
pounds in 2004) and 4700-lb pickups (which weighed 3500 lb in the mid-1980s)
has exacerbated this issue. Good design can mitigate much of this, but matching
two cars of vastly unequal weight is unfair to the driver of the lighter car. The
momentum transfer p to each vehicle is the same in elastic or inelastic collisions.
Thus, the velocity change in a collision is inversely proportional to vehicle mass
(v = p/M) giving,
vlight /vheavy = Mheavy /Mlight . (15.53)
A heavy car with twice the mass of a light car experiences 50% of the change in
velocity (and 50% the force and deceleration) as compared to the light car. Colli-
sions are inelastic, often leaving the cars essentially attached. The velocity change
for a totally inelastic collision between equal masses is vboth = vo . For a mass
ratio 2/1 between cars with the same initial velocity vo in a totally inelastic col-
lision, the light-car velocity change is vlight = 1.33vo and the heavy car velocity
change is vheavy = 0.67vo . This becomes more dramatic for a collision with a
hummer with mass ratio 3/1, the light-car velocity change is vlight = 1.5vo and
the heavy car velocity change is vheavy = 0.5vo . These results are consistent with
fatality data that shows a 2.3 times higher risk for light car passengers in collisions
when Mheavy /Mlight = 1.6. The safety problem would be worse if collisions were
(nonexistent) elastic collisions, a situation that would give larger velocity changes.
An elastic collision between cars of equal mass at equal speeds would give each
car a velocity change v = 2vo , twice that of a totally inelastic collision.
Not surprisingly, the data show there are many fewer deaths from collisions of cars
with equal masses compared to the number that results from collisions between
cars with different masses. In 2002 the NAS concluded that a 10% weight reduc-
tion of CAFE cars (1400 kg) costs 850 fatalities/year, while a similar reduction in
SUVs/light-trucks (2000 kg) saves on average 350 lives. The NAS further concluded
that a 10% weight reduction of all autos/SUVs increases fuel economy by 7%, sav-
ing 0.5 Mbbl/day. Many lives could be saved with a mandated limit on vehicle
mass, one that required the same car mass for everybody.
15.7 Transportation Potpourri

15.7.1 Freeway Noise
Acoustical noise power emitted by cars at highway speeds is about 0.01 W. If the
spacing between cars is 20 m, the acoustical noise from a 20-m length of 10-lane freeway
P1: GFZ
15.7. Transportation Potpourri 397
cylinder is
P20 = (10 lanes)(0.01 W/car) = 0.1 W. (15.54)
Sound from one auto is a point source that decreases inversely with square of the
distance, but sound from a line of traffic decreases inversely with a single factor of
distance. Sound intensity I is obtained by balancing the power emitted by cars in
an arbitrary length L = 20 m to the power passing through a cylinder at radius r ,
I = P20 /2πrL = (0.1 W)/(2π )(30 m)(20 m) = 0.27 × 10−4 W/m2 , (15.55)
where r = 30 m. The corresponding decibel level is significant at
dB = 10 log[I /Io ] = 10 log[0.27 × 10−4 W/m2 /10−12 W/m2 ] = 74 dB, (15.56)
with Io = 10−12 W/m2 . If the distance is doubled to r = 60 m, the intensity is cut

in half, but the sound level is reduced only 3 dB to 71 dB. If the distance is doubled
again to 120 m, a second 3-dB reduction lowers the level to 68 dB. Reduction due
to air absorption of 7 dB/km is more significant at greater distances, but at 120 m
it causes an additional drop of only 1 dB. The slow drop-off of freeway noise is the
reason freeways are shielded with high barriers.
15.7.2 Jet Noise

A jet aircraft emits 105 acoustic watts into a spherical area 4πr 2 . This gives a painful
intensity at 100 m of
I = 105 W/(4π )(102 m)2 = 1 W/m2 , (15.57)
producing an ear-shattering 120 dB. Geometrical spreading reduces intensity to

100 dB at 1 km and to 80 dB at 10 km. Air absorption has a bigger effect, reducing
the level by 7 dB/km to 10 dB at 10 km. Geometrical spreading predominates when
noise is close, while absorption predominates when noise is distant. The average
noise at Los Angeles International Airport has been as high as 85 dB.
15.7.3 Airline Energy

From 1970 to 2001, commercial aviation energy intensity dropped from 10,400 to
3900 Btu/passenger mi. These averages are larger than that of a modern, fully loaded
Boeing-777, which gets 1600 Btu/seat mi. A Boeing-747 burns kerosene at a rate of
110 MW. In 1 h the 747 travels 500 mi, burning fuel at the rate of
(110 MW)(3600 s/h)(1 gal/130 MJ) = 3000 gal/h = 10 tons/h. (15.58)
The 747 fuel economy is 500 mi/3000 gal = 0.17 mpg. The 350-seat plane has a
passenger fuel economy of
(350 passengers)(0.17 mpg) = 60 seat mi/gal, (15.59)

P1: GFZ
twice that of a single-occupied CAFÉ compliant car. The 747 has an energy use
intensiveness of
(130 MJ/gal)(1 gal/60 seat mi) = 2.2 MJ/seat mi = 2100 Btu/seat mi. (15.60)
For comparison sake, the older 707 got 3300 Btu/seat mi and the newer 777 gets
1600 BTU/seat mi. In 2001, US airlines flew 1.1 × 1012 seat mi with average energy
intensiveness 50% greater than the 747, consuming
E airlines = (1.1 × 1012 seat mi)(3.3 MJ/seat mi) = 3.6 × 1018 J

= 3.4 quads = 1.6 Mbbl/day. (15.61)
A 747 round trip from Washington to Paris consumes fuel for each passenger of
(8000 mi)(1 gal/60 mi) = 130 gal jet fuel, (15.62)
which at $2/gal is about 25–50% the cost of an economy ticket.
15.7.4 Freight Modal Shift

Rail freight energy intensiveness at 400 Btu/ton mi is much lower than energy
intensiveness of heavy trucks at 3400 Btu/ton mi. If 20% of US truck shipping’s
7.9 × 1011 ton mi (2001) were switched to trains, the energy saved would be
(0.2)(7.9 × 1011 ton mi)(3400 − 400)(Btu/ton mi)(1 quad/1015 Btu)

= 0.5 quad/year. (15.63)
This modal shift would save 0.5% of US energy. The route options of trucks
are more flexible than trains, but speed and reliability also matter. Unfortunately,
a truck with a single driver carries lettuce from California to Manhattan faster
and more reliably than trains, which is why a rancher friend uses trucks. Perhaps
levitated magnetic trains will be the next wave of trains. In 2003, China agreed to
let Germany extend its 30-km levitated train from Shanghai by 300 km to nearby
cities. The train reportedly goes 425 km/h (266 mph).
Problems
15.1 Drag coefficient. Measure the time it takes your car to slow from 60 mph
to 55 mph in both directions on a flat road. Determine its aerodynamic drag
coefficient Cd and rolling friction coefficient Cr . What is its drag power at 60
mph?
15.2 Down-hill versus drag. (a) Determine the slope of a hill that gives a constant
coasting velocity of 30 mph for a car with Cd = 0.3, Cr = 0.01, mass 1000 kg,
area 2 m2 . (b) What is air drag power, rolling resistance power, and gravity
power?
P1: GFZ
Problems 399
15.3 Uphill. Now let the car go up the hill of problem 15.2 at 30 mph. What is the
power needed at the wheels? What is the thermal input power needed for a
17% efficient car?
15.4 CAFE standards. (a) A manufacturer makes 1 million cars with 20 mpg fuel
economy. How many cars must be produced at 35 mpg fuel economy to
comply with CAFE’s 27.5-mpg standard? (b) How much oil (Mbbl/day) will
be saved when the United States increases the standard for SUVs/light trucks
from 20.7 to 22.2 mpg (2007), and then to 27.5 mpg? (c) The United States has
74 million SUVs/light trucks that travel 12,000 mi/year. How much money
is saved if gasoline costs $2.50/gal? (d) How much oil and money is saved if
CAFE is raised on 140 million cars from 27.5 to 32 mpg; to 40 mpg? (e) How
much oil, carbon (2.4 kg C/gal) and money would be saved if all 200 million
cars/light trucks/SUVs obtained 32 mpg; 40 mpg; and 80 mpg?
15.5 Feebates. (a) Assume a revenue-neutral feebate program (the government
breaks even) for two cars that sell in equal numbers each year; SUVs at
20.7 mpg ($30,000 price), and economy cars at 40 mpg ($20,000 price). Use
the fleet-average fuel economy as the balance point for feebates. Assuming a
penalty of $500/mpg for the SUV, what should the rebate to the fuel economy
car owner be to result in neutral revenue for the government? (b) How will
sales of the two cars change with market forces, assuming demand elasticity
e d = −0.1 and −0.2? What is the new balance point after market forces have
acted, and how much gasoline is saved for a fleet of 200 million? (c) How
much does the new balance point affect revenue flow?
15.6 Otto cycle. Derive Eq. 15.27 for the efficiency of the Otto cycle.
15.7 Diesel cycle. Derive an equation for the efficiency of the Diesel cycle.
15.8 Otto versus Diesel. Gas at 300 K is compressed by a factor of 10 in an Otto-
cycle engine and a factor of 15 in a diesel engine. What are the temperatures
and pressures for these two situations? What are the 50% Carnot efficiencies
for the two engines? What are the Otto and Diesel 50% efficiencies from
problems 15.6–7?
15.9 Extra molecules. (a) Redo problem 15.6, correcting for the fact that burn-
ing octane converts 13.5 molecules into 17 molecules. How much does this
raise the Otto efficiency? (b) How many eV/atom are released from heptane
(48 kJ/g)?
15.10 Regenerative breaking. Describe a scenario for suburban driving and deter-
mine the fraction of energy saved with regenerative breaking.
15.11 Flywheel storage. Determine the flywheel mass needed to recover energy of
a 1500-kg car at 60 mph. Assume the solid cylinder with a 20-cm diameter
rotates at 60,000 rpm.
15.12 Cost of conserved energy. A car is driven 12,000 mi/year at 28 mpg at
$2.50/gal. What is the break-even investment for a car that gets 80 mpg
at with a capital recovery rate of 12%/year?
15.13 Hydrogen airplanes. In the 1970s NASA wanted to convert two large jets
from petroleum fuel to liquid hydrogen fuel. How much hydrogen would be
needed to fly a 747 across the country if it burns kerosene at 110 MW?
P1: GFZ
15.14 Airplane drag. What is the drag force on a 747 traveling at 600 mph at a
10-km elevation where air density drops to 0.5 kg/m3 ? Assume Cd = 0.3
and area = 75 m2 . How does drag power compare to fuel consumption of
110 MW?
15.15 3 dB. Show that halving sound intensity lowers decibel level by 3 dB. What
distances are needed to halve sound intensity for (a) a point source and (b) a
linear source 100 m away?
15.16 Safety costs. Estimate the annual cost to the United States from traffic acci-
dents assuming the following: 40,000 fatalities/year, 100,000 severe medical
handicaps per year, 1 fatality per 300 accidents, 1.5 million accidents a year at
$1000 each, cost of a car is $15,000, worth of a life is $5 M. Go to www.bts.gov
for actual values.
15.17 BART. Should the top speed of San Francisco’s Bay Area Rapid Transit
(BART) be raised from 80 mph to 120 mph? BART’s maximum acceleration
is 2.5 mph/s with 20 s in stations that are 2 mi apart.
15.18 Truck vs. train. How much energy would be saved by shifting 300 billion
ton mi of freight from trucks to trains? Railroads consume 400 Btu/ton-mi
while trucks consume 3000 Btu/ton mi.
15.19 Time horizon. A 2002 NAS report concludes that a 3-year payback limit for
SUV break-even fuel economy happens at 24 mpg, while a 14-year payback
raises the break-even fuel economy to 28 mpg. Explain this difference (see
Chapter 16).
15.20 Parallel or series hybrid. What are the differences between a hybrid auto with
an electric motor in series and one with its motor and engine in parallel?
15.21 Mass versus safety. (a) Analyze the NAS conclusion that a 10% weight
reduction of CAFE cars (1400 kg) costs 850 fatalities/year, while a similar
reduction in SUVs/light trucks (2000 kg) saves 350 lives. (b) Also, analyze
the NAS conclusion that a 10% weight reduction of all autos/SUVs increases
fuel economy by 7%, saving 0.5 Mbbl/day.
15.22 Crush space. (a) What is the crush distance for each car of equal mass at 65
mph with a crush time 0.1 s? (b) What is the crush distance for each car if the
mass ratio is 2?
Bibliography
Davis, W., M. Levine and K. Train (1995). Effects of feebates on vehicle fuel economy, carbon
emissions and consumer surplus, DOE/PO-0031, Department of Energy, Office of Policy,
Washington, DC.
DeCicco, J. (1997). Transportation Energy and Environment, American Council for an Energy
Efficient Economy, Washington, DC
DeCicco, J., J. Kliesch and M. Tomas (2001). The Environmental Guide to Cars and Trucks,
American Council for Energy Efficient Economy, Washington, DC.
DeCicco, J. and D. Gordon (1993). Steering with Prices: Fuel and Vehicle Taxation as Market Incen-
tives for Higher Fuel Economy, American Council for Energy Efficient Economy, Washington,
DC.
P1: GFZ
Bibliography 401
Department of Transportation (2004). National Transportation Statistics, BTS/DOT, Wash-

ington, DC.
Energy Information Administration (2004). Annual Energy Outlook, EIA, Washington, DC.
Friedman, D., J. Mark, P. Monahan, et al. (2001). Drilling in Detroit, Union of Concerned
Scientists, Cambridge, MA.
Greene, D. and J. DeCicco (2000). Engineering-economic analysis of automobile fuel economy
potential, Ann. Rev. Energy Environ. 25, 477–535.
Lee, J. (2000). Historical and future tends in aircraft performance, cost and emissions, Ann.
National Research Council (1992). Automobile Fuel Economy: How Low Can We Go, National
——— (2000). Review of Research Program of the Present Generation of New Vehicles, National
——— (2002). Effectiveness and Impacts of CAFE Standards, National Academy Press,
Washington, DC.
Office of Technology Assessment (1991). Improving Automobile Fuel Economy, OTA,
Washington, DC.
——— (1993). Energy Efficiency Challenges and Opportunities for Electric Utilities, OTA,
Washington, DC.
——— (1995). Advanced Automobile Technology, OTA, Washington, DC.
Ogden, J. (1999). Prospects for building a hydrogen energy infrastructure, Ann. Rev. Energy
Environ. 24, 227–279.
Rennie, J. (Ed.) (1997). The Future of Transportation, Sci. Am. 277(4), 54–137.
Ross, M. (1994). Automobile fuel consumption and emissions, Ann. Rev. Energy Environ. 19,
75–112.
Ross, M., D. Patel and T. Wenzel (2006). Vehicle design and the physics of safety, Phys. Today
57(1), 49–54.
Srinivasan, S., R. Mosdale, P. Stevens, et al. (1999). Fuel cells, Ann. Rev. Energy Environ. 24,
281–328.
P1: GFZ
16
Energy Economics
“. . . there is scarce perhaps a single instant in which any man is so perfectly and
completely satisfied with his situation as to be without any wish of alteration
or improvement of any kind . . . He . . . neither intends to promote the public in-
terest, nor knows how much he is promoting it . . . he intends only his own gain,
and he is in this, as in many other cases, led by an invisible hand to promote an
end which was no part of his intention.”
[Adam Smith, The Wealth of Nations, 1776]
“Famine seems to be the last, the most dreadful resource of nature. The power
of population is so superior to the power of the earth to produce subsistence
for man, that premature death must in some shape or other visit the human
race...gigantic, invisible famine stalks in the rear, and with one mighty blow,
levels the population with the food of the world.”
[Thomas Malthus, An Essay on Population, 1798]
16.1 Basic Economics

The quotes on the “dismal science” of economics by early theorists, Adam Smith
and Thomas Malthus, are relevant today. Thus far, Smith’s vision leads in the de-
bate. Malthus’s prediction has been headed off, somewhat, as new technologies
have increased our capacity to support increased populations. The powers of the
marketplace rode supreme in both Eastern and Western Europe as Marx was buried
a second time, and it continues to do so, however imperfect it is. Energy markets
have had dramatic failures in such things as the lack of progress on energy-efficient
cars and household appliances, until regulations were introduced. Smith’s hidden
hand corrects inefficiencies, but consumers can be apathetic when casually purchas-
ing energy technologies. Hopefully, Malthus’s frightening theory can be buried
along side Marx, but long-term questions on Earth’s sustainability persist when seen
against a backdrop of increasing global consumption.
The economics of energy is all about choices. There is usually a cost of doing
nothing. Optimal choices and opportunity costs can be determined from supply-and-
demand models, but human behavior and congressional laws can tilt the playing
402
P1: GFZ
16.1. Basic Economics 403
Figure 16.1. Price affects petroleum imports. US petroleum supply, consumption, and im-
ports during 1970–2025 in millions of barrels per day. Note that consumption is more reduced
by price rises than domestic production increased by the same price rises. [Energy Informa-
tion Administration (EIA), 2004]
field, favoring one choice over another. While songwriters Erb and Kander told us
in the movie Cabaret that “money, money, money makes the world go ‘round,” we
suggest in this chapter that “energy and money make the world go ’round.”
Today’s cars, houses, and refrigerators consume one-half the energy of their pre-
decessors, and future versions can reduce present consumption by another factor
of two. The excesses of the past were driven by short-term goals of minimizing
cost and maximizing profit. Long-term energy requirements were not a primary
consideration. Today’s appliance regulations direct our choices toward minimized
life-cycle costs that take into account all the costs over a product’s lifetime. See
Fig. 16.1 for the impact of price on petroleum imports and Fig. 16.2 for the U.S.
trade balance.
16.1.1 Compound Interest

Simple interest increases a bank balance linearly with time as C = C0 (1 + it), where
C is the balance after t years, C0 is the initial balance, and i is interest (fraction/year).
On the other hand, the balance with annual compounding is larger,
C = C0 (1 + i)t . (16.1)
For the general case of n compounding payments in a year, the balance after t years
is
C = C0 (1 + i/n)nt . (16.2)
The mathematics is simplified if the balance is compounded continually, that

is, adding interest on a second-by-second basis. The solution to the differential
P1: GFZ
404 16. Energy Economics
Figure 16.2. Trade balance, 1974–2003. Energy imports were 25% of the 2001 US trade deficit
(EIA, 2003).
equation
dC/dt = iC, (16.3)
by integrating over time gives
C = C0 e it . (16.4)
Thus, continuous compounding gives better mathematical models because it gives

rise to differential equations whose solutions are continuous functions. If interest
rates are very small and are applied over small periods of time (it 1), the annual
and continuous compounding approaches are essentially the same, as the Taylor
expansions indicate:
[annual] C0 (1 + i)t ∼
= C0 (1 + it + t(t − 1)i 2 /2 + . . .) (16.5)
[continuous] C0 e it ∼
= Co (1 + it + i 2 t 2 /2 + . . .). (16.6)
The two approaches are almost resolved if interest rates are slightly adjusted to
give the same annual rate of return. We compare 10% continuous compounding
with 10% compounded at monthly, quarterly, and annual intervals. The balance
P1: GFZ
from continuous compounding of interest on $1 after 1 year is
Ccontinuous = $1e (0.1/year )(1 year ) = $1.1052, (16.7)
which is slightly larger than an account that receives monthly interest payments,
Cmonthly = $1(1 + 0.10/12)12 = $1.1047. (16.8)
The monthly and continuous balances are larger than balances that are paid over
longer periods of time, such as quarterly payments ($1.1038) and one annual pay-
ment ($1.10). Although the differences seem small for single-year calculations, large
projects are often funded over 30 years, which makes large differences between
types of compounding. A $1 investment over 30 years at 10% continuous interest
grows to $20.09, similar to the return from monthly payments ($19.84), which are
both larger than the return from annual payments ($17.45).
16.1.2 Buying Bonds

Using the exponential formula, we can determine the interest rate used for a $3000
bond that matures to $4000 in 4 years.
Cfinal = $4000 = C0 e it = $3000e 4i (16.9)

ln(4000/3000) = 0.288 = 4i, or i = 7.2%/year. (16.10)
16.1.3 Law of 70
It is useful to use the idea of doubling times for the purpose of making quick
estimates. An investment that has exponential growth is doubled according to
2C0 = C0 e i T2 . (16.11)
This is solved for doubling time T2 by taking the natural log of both sides and
shifting interest from fraction/year to percent/year, to obtain the law of 70:
T2 = 70/i (16.12)
with T2 in years and i in %/year. A caution: this formula is inaccurate for very
high interest rates. For example, an investment at 100% per year interest rate has a
doubling time of 1 year, while the formula gives T2 = 0.7 year. Usual interest rates
give much less error. The law of 70 can be applied to other kinds of growth besides
monetary. Prior to the oil embargo, US electrical power was growing at 7%/year,
a rate that would have required the power grid to double every 10 years (not
counting replacement of old power plants). Thirty years of 7% growth would be
three doublings, raising the grid size by a factor of 8. But this is not what happened.
Instead electrical growth adjusted to 1.8%/year and energy growth to 1.5 %/year
(EIA, 2003). At these rates the doubling times were 39 and 47 years, respectively.
P1: GFZ
16.1.4 Current and Constant Dollars

There are two types of money: one being current dollars, which includes the effects
of inflation, and the other, constant dollars, which removes the effects of inflation.
Banks pay interest and list balances in terms of current dollars, that is, using present
monetary value. The interest rate in current dollars is typically the inflation rate
plus the true cost of money, which is about 3%/year.1 This approach takes its lead
from the trend of interest rates after the oil shocks of the late 1970s, when the rates
rose to 18% while the inflation was 15%. Future-based calculations are often done
in constant dollars because future inflation rates are unknown.
16.1.5 George Washington’s Dollar

On December 25, 1776, George Washington (supposedly) threw a silver dollar
across the Delaware River. What would be the value of Washington’s silver dollar if
he had invested it rather than thrown it? If the average rate of inflation is 2%/year,
then the cost of money in current dollars would perhaps be 5%/year. This rate would
give Washington’s heirs a bank balance in 2004 in current (real) dollars of
C5%−2004$ = $1e (0.05/year)(228 year) = $1e 11.4 = $89,000. (16.13)
The account would have a balance of about $9100 at 4% and a balance of $830,000
at 6%. These dollars are relevant in 2004, but President Washington would have
deemed these figures too high since the 2004 dollars are less valuable. We accom-
modate this view by calculating the value of the $1 investment in 1776 constant
dollars at 3%/year:
C3%−1776$ = $1e (0.03/year)(228) = $1e 6.84 = $930. (16.14)
At 2% the $1 investment is worth only $96 in 1776 dollars.
16.1.6 Fixed Payment Loans

Most loans are paid periodically with fixed payments that cover current interest and
reduce the unpaid balance. The first payments mostly pay interest on the balance, that
is, iC. The last payments mostly lower the balance, that is, dC/dt. These two effects
are additive in determining the fixed payment rate P, which is the capital recovery rate
α times initial investment: P = αC0 . This gives
P = αC0 = iC − dC /dt. (16.15)
1
Even without inflation, money has historically had an income rate since there is an oppor-
tunity cost that one forgoes by not investing. The banks pay customers the historical cost rate
(plus inflation rate), allowing the banks to make money on their investments, such as real
estate. Of course, there are times when the true cost of investing money is negative. This
occurs when market values drop.
P1: GFZ
The capital recovery rate is also known as the annual fixed payment rate. The
formula integrates to give the time to reduce the balance to C:

1 − iC/P
t = (1/i) ln , (16.16)
1 − i/α
which, solved for C, gives the balance as a function of time,
C = C0 (α/i)[1 − (1 − i/α)e it ]. (16.17)
Time needed to pay the entire loan (i.e., C = 0 and t = T) is

1
T = (1/i) ln . (16.18)
1 − i/α
For an investment that is to be paid in full over T = 30 years at i = 6%/year, the
annual fixed payment rate would be
α = i/(1 − e −i T ) = 0.06/(1 − e (−0.06/year)(30 year) ) = 0.06/(0.835) = 7.2%/year.
(16.19)
This agrees with published tables: For example, a $100,000 loan at 6% is listed with
a payment of $600/month or $7,200/year (α = 7.2%/year). Over the 30 years the
investor pays 7.2%/year × 30 years, or 220% of the project, of which 55% goes for
interest and 45% for principle. An interesting thing happens at higher interest rates.
For comparison, a 9% loan requires α = 9.66%/year to pay the debt, a figure that
is closer to the 9% interest rate than the 7.2% payment was to the 6% interest rate.
This is not a benefit to the borrower. Higher interest rates dedicate more money
for interest and less for debt reduction. This is dramatically seen in the case of 18%
interest, which gives a payment rate α = 18.1%/year. The point is that the higher
interest rates give capital recovery rates that are very close to the interest rate.
16.1.7 Supply/Demand Elasticity

Previous chapters used the concepts of supply elasticity and demand elasticity to
explain policies on carbon (Section 8.8) and gasoline consumption (Section 10.3).
For example, we saw higher gasoline prices encourage drivers to purchase less.
To review, demand elasticity couples fractional change in price (p/ p) to fractional
change in demand (d/d):
d/d
ed = . (16.20)
p/ p
Elasticity is determined at the price where supply and demand are matched and it
will be less accurate to either side of that equilibrium. Demand elasticity is negative,
since p, a rise in price, reduces demand d. The value of e d is hotly debated and
changes with time, place, and conditions. In a totally elastic market, a rise in price
of x% reduces demand by x%, giving e d = −1. In a totally inelastic market, a rise
in price does not change demand, giving e d = 0. Most markets lie between these
extremes.
P1: GFZ
Supply elasticity is the same type of function as demand elasticity:

s/s
es = , (16.21)
p/ p
where s is supply. The critical difference from demand elasticity is that supply
elasticity is positive, since a rise in price encourages an increased supply through
more production.
16.1.8 Gasoline Demand

Gasoline is closer to the inelastic limit, since driving habits change only slightly
when gas prices change a small amount. However, elasticity is not constant; it
increases with larger price rises. For instance, a 1 c/gal rise in price has essentially
no effect, while a $1/gal rise has an effect more than 100 times the effect of the 1 c/gal
rise. Elasticity can rise over time: For example, it may increase as consumers have
the opportunity to consider buying smaller cars and moving closer to work. But it
can also drop over time, for example, as consumers adapt to higher prices. Thus,
demand elasticity can be regarded as a kind of psychological “spring,” one that
depends on human nature and market forces.
Using demand elasticity, we estimate the effect of a $2/gal conservation tax on
US gasoline on top of the present cost of about $3/gal. Such a tax would bring
US prices closer to European prices of $5/gal. We assume a small elasticity of
e d = −0.1 in the near term and a greater one in the long term, say e d = −0.2, as
some consumers would change travel patterns after a long period of higher prices.
Long-term reduction of US gasoline consumption of 8.7 Mbbl/day (2001) would
be
d = e d d(p/ p) = −0.2(9 Mbbl/day)($2.00/$3.00) = −1.2 Mbbl/day, (16.22)
which is a 13% reduction in consumption to 7.5 Mbbl/day. Besides conserving
petroleum, the conservation tax could raise considerable revenue, namely
($2/gal)(42 gal/bbl)(6.4 Mbbl/day)(365 day/year) = $200 billion/year, (16.23)
which could be used to reduce federal taxes by 20%. The conservation tax is an
unpopular notion among most US consumers.
16.1.9 Cross Elasticity

Gasoline demand would be reduced if the price of a second fuel, such as natural
gas, were reduced. The cross-supply demand price function is
d1 = K p e11 e12
1 p2 , (16.24)
where d1 is demand of gasoline, p1 and p2 are prices of gasoline and natural gas,
respectively, and K is a constant. The self-elasticity e 11 links gasoline prices to gaso-
line demand, while cross-elasticity e 12 links natural gas prices to gasoline demand.
P1: GFZ
16.2. Discounted Benefits and Payback Periods 409
Taking the differential with gasoline price p1 held constant gives the cross-elasticity,
e 12 = (d1 /d1 )/(p2 / p2 ). (16.25)
A hypothetical 10% drop in natural gas price that decreases gasoline demand by
1% gives e 12 = 0.1.
16.2 Discounted Benefits and Payback Periods

Businesses hesitate to make investments with payback periods over 5 years (a fig-
ure that corresponds to 14%/year interest, as computed from 70 = λT2 ). Electrical
utilities accept longer payback periods because they know public utility commis-
sions remove risks by guaranteeing that consumers will pay for investments in the
rate structure. This allows utilities to get favorable loan rates. On the other hand,
agricultural firms usually restrict investments to a payback of one year since they
have few protections and they have large cash flow needs. Homeowners who buy
solar hot water heaters or increase insulation instinctively accept a longer payback
of 10–20 years. The purchasing of energy-efficient refrigerators is particularly at-
tractive with a payback of 1–2 years (Section 14.5).
16.2.1 Discount Rates and Discounted Benefits

Proponents of large projects sometimes make misleading statements in which they
quote benefits of a project by quoting benefits far into the future. This is often done
with the understanding that the public may not know that future benefits are less
valuable when translated into the present. To be sure, there is an opportunity cost
to not investing. However, the most honest way to quote benefits is to state them
in terms of the present. This is done using a function that determines present value
(PV) of an investment after a number of years. The “present” refers to the time when
a decision is made. Present value calculations remove the vagaries of inflation. As
an example, having $1000 today is preferable to having $1000 in the future, because
today’s $1000 can be invested to obtain dividends starting now. Tomorrow’s $1000
benefit is worth less than $1000 in the present because it has not acquired dividends
over time. Thus, investors should discount future benefits (returns) by a procedure
called discounting. The amount of discounting depends on the future date of the
benefit and the discount rate d. Capital Co invested at an interest rate i has a future
value
C = C0 e it . (16.26)
To consider future benefits in the present, we reverse the process shown in this
familiar formula by using a discount rate close to the interest rate (d = i) and run
the clock backward with a negative coefficient on the time term. Bfuture , benefit
obtained t years in the future, is discounted continuously to give its present value
P1: GFZ
PVB via
PVB = Bfuture e −dt . (16.27)
Reverse annual compounding can also be used for discounting. The calculation is
given by
PVB = Bfuture (1 + d)−t , (16.28)
which gives the same result as the calculation for continuous compounding in the
case of small discount rates and short times. The choice of discount rate is extremely
important. It can be misleading if discount rate is determined to be equal to present
interest rate in a given situation. Interest rate is approximately the sum of the cost
of money, perhaps 3%/year, and an estimate of future inflation, perhaps 2%/year.
The value of saving energy rises if inflation causes future energy prices to go up.
But since other investments also benefit from inflation (that is, the benefit to energy
saving is relative), the removal of inflation from the discount rate seems reasonable. On
the other hand investors take risks, a fact that tends to introduce the need to raise
discount rates. The value of a present benefit in the future, one that is driven only
by inflation, is
Bfuture = PVB e λt , (16.29)
where λ is the inflation rate. Hence, to obtain the present value of the benefit without
inflation, the future must be discounted to the present as
PVB = (Bfuture eλt )(e −dt ) = Bfuture e −(d−λ)t . (16.30)

The reduced discount rate (d − λ) is the discount rate in current dollars. Using
d = i, leaving out λ completely, would make the discount rate higher and would
discount future benefits more. Lowering the discount rate to (d − λ) to remove
inflation reduces future benefits less, making investments appear wiser. The courts
in 1985 denied DOE’s “no-standard standard” that tried to negate appliance energy
standards. In summary, for our calculations we prefer a discount rate of 5%/year
for energy conservation measures. This removes inflation’s 2%/year, replacing it
with the same percent to take into account risk and maintenance.
16.2.2 Simple Payback Period

Simple payback period is the time it takes to pay back an original investment in
constant dollars, without considering interest or inflation. An energy saving device
costs C0 = $1000 with benefit rate RB = $200/year in reduced fuel costs. It has a
simple payback period T, which is the ratio of investment cost divided by savings
per year,
T = C0 /RB = $1000/$200/year = 5 year. (16.31)
If the interest is small, the capital is paid off almost linearly, giving average
interest on one-half the investment. An interest of i = 5%/year gives an approximate
P1: GFZ
16.2. Discounted Benefits and Payback Periods 411
average interest cost of ($1000/2)(0.05) = $25/year, which neglects compounding.

This reduces the benefit to $175/year, increasing T to $1000/$175 = 5.7 year.
16.2.3 Discounted Payback Period

The simple payback approach does not discount future benefits. Using annual
compounding, the payback period T is determined by summing the discounted
benefits until the total present value equals the $1000 investment cost, or PVB = C0 .
Using d = 5%/year, this gives
PVB = RB (1 + d)−t (16.32)
$1000 = $200[(1.05)−1 + (1.05)−2 + · · · + (1.05)−T ], (16.33)
which is satisfied for T = 6 year. It took longer to regain the $1000 investment
with discounted benefits than with a simple payback because compounded future
benefits are less valuable in the present.
Annual compounding has limitations for long payback periods because it in-
volves summing many terms. A better approach in such cases is to integrate contin-
uous future benefits until year T when the present value is equal to the investment
cost C0 .
T
PVB = C0 = RB e −dt dt = (RB /d)(1 − e −dT ). (16.34)
0
(Important to note: d in the exponent is the discount rate and not to be confused
with the differential dt.) We have assumed that discount rate d is in constant dollars,
without inflation but with risk and maintenance. If fuel costs rise at a rate φ above
the inflation rate λ, the present value of benefits over T years is
PVB = [RB /(d − φ)](1 − e −(d−φ)T ), (16.35)
Taking the natural log with φ = 0 gives,
T = −(1/d) ln[(RB − dC 0 )/RB ], (16.36)
which gives T = 5.8 year. For small discount rates, a Taylor expansion of Eq. 16.36
with ln(1 − x) ∼
= −x gives the simple payback period, T = C0 /RB .
The highest total of discounted benefits comes after infinity years,
PVB = [RB /d](1 − e −(d−λ)T ) = (RB /d) = $200/0.05 = $4000. (16.37)
After 30 years PVB is almost as high, at $3100, as all benefits after 30 years are worth
only $900 in the present.
The net present value is the present value of all benefits after a given number of
years minus the capital cost (PVnet = PVB − C0 ). For the case of T = 30 year, PVnet =
$3100 − $1000 = $2100. An investment can never be recovered if the annual interest
on it equals the annual benefit. For example, if capital cost in Eq. 16.36 had been four
times higher at $4000, with interest of $200/year, the payback would be T = ∞. If
future fuel prices rise above the inflation rate, the investment picture brightens. If fuel
P1: GFZ
inflation φ = 2.5%/year is included, payback period is reduced from T = 5.8 year

to 5.4 year (C0 = $1000, RB = $200/year, and d = 5%).
Finally, for complete accounting, annual maintenance M and one time repair cost
CR in year T must be subtracted from the present value of future benefits,

PVB = [(RB − M)/(d − φ)](1 − e −(d−φ)T ) − CR e −dT . (16.38)
16.3 Cost of Conserved Energy

A useful way to examine cost effectiveness of energy-conserving measures is to
determine the cost of conserved energy on the investment. If the cost of saving a kilo-
watthour is less than the local cost of a kilowatthour, the measure is cost-effective
and it probably is worth doing. For an investment of about $150, refrigerators in
the 1970s saved 1300 kWh/year, giving a simple payback period of
T = cost/annual saving = C0 /RB
= $150/(1300 kWh/year)($0.085/kWh) = 1.4 year (16.39)
with electricity at $0.085/kWh. The cost of conserved energy CCE is a better mea-
sure of cost effectiveness than length of payback period because it can be directly
compared to local energy costs. CCE is defined as
CCE = annual cost/annual energy savings = (CRR)(C0 )/RE , (16.40)
where Co is investment cost and RE annual energy savings. The capital recovery rate
CRR is the payment rate per dollar of investment required to pay the investment
and the interest. The CRR is α in Eq. 16.19, which gives the annual payment formula
P = CRR × C0 = iC o /(1 − e −i T ). This gives the capital recovery rate,
CRR = P/C0 = i/(1 − e −i T ). (16.41)
For annual compounding, the formula is
CRR = i/(1 − (1 + i)−T ). (16.42)
The capital recovery rate for a 20-year loan with 5% continuous compounding
in constant dollars is
CRR = 0.05/(1 − e −(0.05/year)(20 year) ) = 7.9%/year, (16.43)
while for annual compounding it is slightly higher,
CRR = 0.05/[1 − (1.05)−20 ] = 8.0%/year. (16.44)
At an investment of $150, refrigerators save 1300 kWh/year over older models.
Calculating the CCE shows that this is an excellent monetary investment with
CCE = (CRR)(C0 /RE ) = (0.08/year)[$150/1300 kWh/year] = 0.9 c kWh, (16.45)
which is much less than residential electricity at 9 c/kWh (2004). The conservation
supply curves in Figs. 16.3 and 16.4 give the total energy savings as a function of
P1: GFZ
16.3. Cost of Conserved Energy 413
Figure 16.3. Building sector cost of conserved energy. The Five-Lab DOE study predicted a
16% drop in electricity use in the building sector if new technologies were implemented in
65% of these cost-effective situations: (1) commercial lighting, (2) commercial heating and
air conditioning, (3) commercial refrigeration, (4) residential lighting, (5) residential heating
and air conditioning, (6) commercial water heating, (7) commercial other uses, (8) residential
refrigerators and freezers, (9) residential water heating, (10) residential other uses. Savings
from reflective roofing are contained in the residential and commercial space conditioning
categories. The projected savings of 400 TWh/year in 2010 corresponds to 45 GWe plants
operating full time (Brown et al., 1998).
Figure 16.4. Supply curve for residential lighting in California. The supply curve for the cost
of conserved energy (CCE) for residential lighting shows that 20% savings can be obtained
on a cost-effective basis. The conservation measure is economic when the CCE is less than
the price of electricity (Rosenfeld, 1999).
P1: GFZ
CCE. These data show that considerable energy can be saved at CCEs less than the
residential price of electricity.2
16.4 Minimum Life-Cycle Costs

The electricity saved by improving refrigerators from 1800 kWh/year (1974) to 450
kWh/year (2001) will eliminate the need to build some 40–50 1-GWe plants and will
save $16 billion/year (Section 14.5). The economics of this can be brought to light by
considering life-cycle cost, which is the sum of purchase cost, energy costs, and any
other costs. If life-cycle costs are minimized, savings to consumers are maximized.
The pre-embargo marketplace failed to produce intelligent refrigerators, but fear of
another energy crisis convinced the US Congress to require life-cycle costing as the
basis of establishing appliance standards. This section is based on the 1985 work of
Mark Levine and his colleagues at Lawrence Berkeley National Laboratory, which
convinced the courts to establish congressionally mandated, meaningful energy
standards for appliances. LBL estimates that appliance standards saved the United
States $48.5 billion by 2000.
The net present value of the costs and benefits of an appliance is
T
PVnet = (RC − RB ) e −dt dt, (16.46)
0
where RC and RB are the cost and benefit rates. (Again, note that the exponential
term is the product of discount rate and time). Annual energy used (E) depends
on appliance purchase cost C in the following manner,
E = E ∞ + (E 0 − E ∞ ) exp[−A(C/C0 − 1)], (16.47)
where A is a constant, E ∞ is the rate of energy consumption at an infinite purchase
cost, and E 0 is the rate of consumption during the base year at a purchase cost
of C0 . The optimal investment is obtained by minimizing life cycle cost, LCC =
C + PVenergy , over product lifetime. Energy price p increases at a rate φ above
inflation for an annual cost in current dollars of CE = Ep e φt .
The present value of the energy cost over the appliance’s lifetime is
T
1 − exp[−(d − φ)T]
PVE = E p e −(d−φ)t] dt = E p . (16.48)
0 (d − φ)
The life-cycle cost for an appliance is LCC = C + P V E , or
LCC = C0 {1 − ln[(E − E ∞ )/(E 0 − E ∞ )]/A} + PVE . (16.49)
The minimum LCC is obtained by minimizing LCC with respect to E,
[d(LCC)/dE] E min = −(C0 /A)/(E min − E ∞ ) + p[1 − e −(d−φ)T ]/(d − φ) = 0. (16.50)
2
The DOE values are credible, but there are differences between experts (Joskow and Marron,
1993).
P1: GFZ
16.5. Energy Tax Credits 415
This gives the optimal energy consumption E min at the minimum LCC,
E min = E ∞ + (d − φ)(Co )/(Ap)[1 − e −(d−φ)T ]. (16.51)

The minimum value of LCC and corresponding energy use and purchase cost
are obtained using Eqs. 16.49 and 16.51 with Levine’s 1985 parameters: Life-cycle
cost for a refrigerator dropped $503, from base-case LCC of $1704 to its minimum
value of $1201. Energy use dropped 705 kWh/year, from E 0 = 1217 kWh/year
to E min = 512 kWh/year at minimum LCC, which is not too different from E ∞ =
475 kWh/year (the infinite purchase price). The purchase price rose $94, from base-
case C0 = $674 to Cmin = $768 at the minimum LCC. The 1984 calculations used
p = $0.069/kWh, A = 21.6, d = 5%/year, and φ = 0.
The additional capital cost, C, for improved refrigerators
C = Cmin − C0 = p(E o − E min )[1 − e −(d−φ)T ]/(d − φ), (16.52)

is recovered with a payback period of
T = ln[1 − C(d − φ)/ p(E 0 − E min )]/(φ − d) = 2 year. (16.53)
16.5 Energy Tax Credits

16.5.1 Solar
Congress established a time-limited tax credit in 1976 to encourage homeowners
to retrofit their homes with solar hot water heaters. We determine the cost effec-
tiveness of the author’s 1979 purchase of a 42-ft2 solar hot water system. The US
and California governments and the Pacific Gas and Electric Company contributed
$2000 for the purchase of the $3500 unit. If 50% of the average integrated daily solar
flux of 2000 Btu/ft2 day (Section 12.1) is usable, this collector saves
d Q/dt = (42 ft2 )(1000 Btu/ft2 day)(365 day/year) = 15.3 M Btu/year. (16.54)
Natural gas is sold in therms of 100 ft3 containing 0.1 M Btu. The 1979 residential
price was about $3/M Btu. Annual energy savings from the solar collector in 1979
was
RB = (15.3 MBtu)($3/MBtu) = $46/year. (16.55)
In 2001 the residential price was $9.68/MBtu, giving RB = (15.3 MBtu)
($9.68/MBtu) = $148/year. Electricity at 3 c/kWh in 1979 gave a cost
($0.03/kWh)(1 kWh/3.6 MJ)(1055 J/1 Btu) = $8.80/MBtu. (16.56)
for RB = $135/year. The author’s solar unit was not cost effective in competition
with natural gas without tax credits. (It was hoped that tax credits for solar water-
heaters would increase production of solar units and drop costs, but this did not
happen.) However the solar hot water heater did compete against electricity, giving
P1: GFZ
a present value in 1979 for 30 years of electricity savings,
PVB = (RB /d)(1 − e −dt ) = ($135/0.05)(1 − e −(0.05/year)(22year) ) = $2000. (16.57)

With today’s tripled electricity prices, the solar option is the clear economic
preference when compared to electricity. Since 1979, residential electricity prices
rose at 4.4%/year, giving a 2001 residential cost of 8.7 c/kWh for $25.35/MBtu and
RB = $388/year. It was not known in 1979 that the present value in 1979 of solar
benefits (vs. electricity) would be $6500 figured at a net discount rate d = 0.025/year.
16.5.2 Conservation
Congress adopted a time-limited, 15% tax credit in 1976 to stimulate energy savings
measures that had paybacks of less than 10 years. The idea behind tax credits is that
they create new purchases that otherwise would not have happened. But tax credits
also go to those who would have taken the energy-saving measure without the tax
credit. We consider the favorable case of ceiling insulation that costs $200 (1975)
with a payback of 2.3 year to save 6 bbl/year (uninsulated ceiling, oil equivalent).
We determine the cost to the government to save 1 bbl of oil for demand elasticities
e d of 0, −0.5, and −1.
If there are no tax credits, there will be N ceiling insulation jobs per year, cost-
ing $200N/year and saving 6N bbl/year, where N is a few million. The 15% tax
credit lowers the $200 cost to the owner to $170. For zero demand elasticity (that
is, for demand affected by price) there would be no additional sales. The United
States would give $30 to persons who would normally purchase an insulation job
for a total of $30 × 6N = $180N. Since no additional barrels were saved because
of the tax credit, the cost per additional barrel saved is $180N/0 bbl = infinity. For
the case of perfect elasticity of e d = −1, there would be 15% additional sales, or
1.15N sales/year, saving the United States (1.15N × 6 bbl/y) = 6.9N bbl/year. The
additional oil savings from the tax credit is (6.9N − 6N)(bbl/y) = 0.9N bbl/year.
The cost to the United States per additional annual barrel saved c US is the cost to
the United States divided by the number of additional barrels per year saved
from the tax credit:
c US = ($30)(1.15N)/0.9N = $38/bbl-saved-year. (16.58)
and c US = $72 for e d = −0.5. The cost is smallest when the market is most elastic.
These costs may seem high. But the cost should not be compared to the 1975–79
OPEC spot price of $13–30/bbl since the $38 to $72/bbl saves 1 bbl/year for the
life of the house. The present value in 1979 of a 1 bbl/year savings over 100 year with
RB at $20/year (1 bbl/year) is
PV1−bbl/year = (RB /d)(1 − e −dt ) = ($20/0.05)(1 − e −(0.05/year)(100 year) )

= $397/bbl-saved-year. (16.59)
and $367/bbl-saved-year for t = 50 year. Depending on the actual value of elastic-
ity, the long-term benefit seems cost effective since it probably outweighs the $38–72
P1: GFZ
16.6. Petroleum Economy 417
government investment. The entire country gets broader savings on saving energy,
but the consumers have a different point of view. The present value of insulating the
ceiling as a result of savings on delivered oil at $30/bbl is
PVceiling = (RB-ceiling /d)(1 − e −dt )
= (6 bbl × $30/0.05)(1 − e −0.05×50 ) = $3300 (16.60)
over 50 years, which far exceeds the $170–200 cost.
16.6 Petroleum Economy

Petroleum is key to modern society, but America’s dependence on it impacts
on foreign policy and national security, as was manifested in both Gulf Wars
demonstrated. Estimates of petroleum production rates with King Hubbert’s model
(Section 10.5) were obtained without consideration of economic principles in an era
when petroleum price was relatively constant and small. In this section, we develop
the economics of petroleum with supply and demand through elasticity functions
that increase supply and reduce demand as prices rise.
The starting point for our model is the Verhulst equation for production from
finite, fixed petroleum resources absent economic considerations (Eq. 10.21);
dQ/dt = λ(Q − Q2 /Q∞ ) = kQ(Q∞ − Q), (16.61)
where dQ/dt is petroleum production rate, λ is initial production growth rate, Q∞ is
ultimate recoverable resource at a constant production cost and k = λ/Q∞ . When
prices rise, less petroleum is used as companies and individuals seek to reduce
spending. We assume price p rises linearly with time, p = φtp, where φ is the first
year fractional petroleum price rise above inflation and t is time in years. Because the
fuel price rises linearly, the quantity demanded decreases linearly, according to
q new = q + q = q [1 + e d (p/ p)] = q (1 + e d φt). (16.62)
where e d < 0. On the other hand, each price rise provides more money, which
can be used to develop lower quality petroleum resources. When this is done, the
recoverable resource Q∞ increases linearly with time, according to
Q∞-new = Q∞ (1 + e s φt). (16.63)
Modifying the Verhulst equation for the effects of supply and demand economics
gives
dQ/dt = k(1 + e d φt)Q[Q∞ (1 + e s φt) − Q]. (16.64)
For historical reasons we use parameters from 1980, the time shortly after the
second oil shock (Iran) when inflation was 15%/year. In keeping with the times,
the petroleum price is increased linearly at 5%/year (for the first year) above mon-
etary inflation. This rate is not unreasonable with regard to a time when oil was
P1: GFZ
Figure 16.5. Modeling US oil production. US domestic production from 1890 to 1980 is com-
pared to the solution of the Verhulst equation without rising prices (curve 1). The Verhulst
equation is modified to take into account the effects of increased supply and reduced demand
at higher prices in curves 2–5, which assume petroleum prices rise linearly at φ = 5%/year
above inflation. (In a time of high rising prices.) (Same answers result from dividing φ by
n and multiplying e by n.). Curve 2 uses only supply elasticity to increase the resource
(e s = 0.2, e d = 0). Curve 3 adds demand elasticity to reduce demand (e s = 0.2, e d = −0.1).
Curve 4 raises demand elasticity as drivers accommodate to change (e s = 0.2, e d = −0.1,
until 1990 when e d = −0.2 for the long-term). Curve 5 symbolizes a synfuel industry that
begins with normal markets with e s = 0.2 until 2000, then the industry is stimulated with
tax credits, raising e s to 0.4, while e d remains at −0.1.
$30–35/bbl (1981 dollars). Since the equations use a product of inflation and elastic-
ity, the results are the same if the price rise is halved to 2.5%/year with doubled elas-
ticity values. The parameters from Section 10.5 are used in a Runge–Kutta routine:
1980 cumulative oil production of 120.4 Gbbl, Q∞ = 165 Gbbl, kQ∞ = 0.0827/year,
and dQ/dtmax , = 3.32 Gbbl/year in 1970. Curves 2–5 are smoothly joined to curve 1
in Fig. 16.5, which is the graph of the solution to Verhulst equation for time prior
to 1980. The assumptions for curves 2–5 are listed in Fig. 16.5.
The results in Fig. 16.5 are what we expect: Higher e s values increase supply and
larger e d magnitudes reduce demand. Trends in production show a great variation
as a result of varying the e parameters. US production in 2001 follows curve 2, which
predicts production of 2.5 Gbbl/year (6.8 Mbbl/day), close to the actual US total
production of 2.1 Gbbl/year (5.74 Mbbl/day), but somewhat above production
for the lower 48 states at 1.8 Gbbl/year (4.8 Mbbl/day). Curve 2 asymptotically
approaches production of
dQ/dt = Q∞ φe s t = (165 Gbbl)(0.05/year)(0.2)(1)

= 1.65 Gbbl = 4.5 Mbbl/day, (16.65)
which is the rate of pumping from the lower 48 states. A low demand elasticity
P1: GFZ
16.7. Imported Oil, Synfuels, and Gasohol 419
of e d = −0.1, which increases to e d = −0.2 (curve 4), is associated with long-term

adjustments by consumers. Lastly, curve 5 speculates birth of a synfuel industry
as a result of increased supply elasticity e s = 0.4 in 2000, which raises Q∞ to 250
Gbbl by 2050. These results do not predict the future, but they help to understand
the dynamics of petroleum supply and demand. Improved technologies play a
role; horizontal drilling, enhanced recovery with solvent, steam and CO2 injec-
tion, improvements in seismology through enhanced diagnostics and computer
mapping—all are examples of these. But there are limits, as the United States is
not finding new large oil fields on its territory. The “estimates” shown on Fig. 16.5
show that the 20-Mbbl/day demand (a figure that is rising) will not be met with
domestic petroleum sources.
16.7 Imported Oil, Synfuels, and Gasohol

In the 1970s, OPEC was strong. Few analysts predicted it would lose its power
to set prices. Yet, this is what happened as OPEC no longer strictly controls pro-
duction by member states, a fact going back to the time beginning when Iraq and
Iran—needing money to pay for their war—stopped producing oil as directed by
OPEC. New sources of petroleum in Mexico and other places further diminished
OPEC’s role, and new technologies expanded the resource base. Conservation, led
by a doubled auto fuel economy in the United States also had a major effect on
OPEC’s control of prices. These factors undercut OPEC but they only postpone
future supply vulnerabilities.
A key conclusion of Yergin and Stobaugh’s Energy Future: Report of the Energy
Project at the Harvard Business School (1979) followed from their examination of indi-
rect economic effects of imported oil. Demand for additional oil—whether Middle
Eastern or from other reserves—allows OPEC to raise its oil price, an increase that
must be applied to both baseline imports and additional imports, increasing oil’s effec-
tive price at the margin over the posted price. Conversely, less demand for OPEC oil
lowers the OPEC price, even though it is a cartel. This can be seen by the follow-
ing calculation: If the non-OPEC nations (called the NON nations here) import an
increased supply of oil s from OPEC, then the price/barrel will rise from p to
( p + p). The direct cost to the NON nations of importing the additional oil would
be (s)( p + p). In addition, the rise in oil price p would apply to the supplys of
oil that NON nations initially imported, giving an indirect cost to NON nations of
(p)(s). The total cost to the NON nations for importing additional oil is the sum
of these two effects:
Ctotal = Cindirect + Cdirect = (s)( p + p) + (p)(s)

= (s + s)(p) + (s)( p). (16.66)
The rise in price p depends on the amount of additional demand s that is

finally obtained from OPEC, and on the estimate of market forces on OPEC. We
assume an effective supply elasticity for the OPEC cartel, e s = (s/s)/(p/ p), which
P1: GFZ
is rewritten to determine change in price p from change in supply s:

p = ( p/e s )(s/s). (16.67)
In economics, marginal cost is the cost of producing one additional unit of an item,
which depends on the production rate s. In our scenario, the total cost of importing
an additional barrel is the cost at the margin. Hence, the total cost/barrel c total to
the importing nations is obtained by substituting the expression for p into the
total cost Ctotal and dividing by s,
c total = Ctotal /s = p[1 + (1/e s )(1 + s/s)]. (16.68)
However, assume the United States, a NON nation, imported the entire extra
supply s, and not its fractional share. Now the marginal cost of oil to the United
States would be less, since indirect costs would be spread out among nations that
did not import more oil. For this case we obtain
c total-US = Ctotal-US /s = p[1 + (1/e s )(sUS /s + s/s)], (16.69)
where OPEC’s initial supply s = 30 Mbbl/day and US imported oil is sUS =
9 Mbbl/day. The total cost per barrel for these two situations, c total and c total-US ,
are plotted in Fig. 16.6 for e s of 0.5, 1, and 2.
The wide variation of costs c total and c total-US depends primarily on elasticity and
secondarily on the rate of imports. There is a large difference between c total and
Figure 16.6. Indirect cost of imported oil. Total (direct + indirect) cost/barrel of importing
more oil from OPEC is shown as a function of OPEC supply s. The solid lines designate
cost/barrel c total in an environment in which additional supplies are shared proportionally
among non-OPEC nations. The dashed lines designate cost/barrel to the US c total-US in an
environment in which only the United States obtains additional supplies from OPEC. Supply
elasticities e s of 0.5, 1 and 2 are used to give a spectrum of values. The direct cost of OPEC
oil was po = $30/bbl (1981 dollars), but it rose to $35/bbl in October 1981.
P1: GFZ
16.7. Imported Oil, Synfuels, and Gasohol 421
c total-US . If one nation were to obtain all the additional oil, the marginal cost of oil
c total-US would be less since, in this case, indirect costs are borne by other nations.
These effects were smaller in 2000, but the National Academy of Sciences used an
indirect cost of $5/bbl in its 2000 study on automobile efficiency.
16.7.1 Synfuels
Congress created tax credits in 1975 to stimulate a synthetic fuel industry. The vision
was not achieved. It seemed that the estimated synfuel cost per barrel always rose to
be a few dollars above the price of oil. Then the price of oil dropped. Environmental
impacts from synfuel production in the western states would have been substan-
tial; nevertheless, it is worth revisiting as domestic oil production declines. The
stakes are high. The economic penalty of the 1973–74 oil embargo may have been
$100 billion, but a synfuel subsidy of $5/bbl to the Synthetic Fuels Corporation for
its projected production of 2 Mbbl/day by 1992 could entail an annual subsidy of
(2 Mbbl/day)($5/bbl)(365 days/year) = $3.7 billion/year. (16.70)
The 1980 estimate for construction of a 50,000 bbl/day synfuel plant was widely
quoted, but uncertain at $2 billion. The annual capital cost/bbl, using a high capital
recovery factor of 16%/year, synfuel being a high-risk industry, would come to
($2 billion)(0.16/year)(1 year/365 days)(1 day/50, 000 bbl) = $18/bbl. (16.71)
A ton of coal can make about 2.5 bbl of synfuel. This gives a 1980 prorated
carbon cost of ($20/ton)/(2.5 bbl/ton) = $8/bbl, to which must be added $10/bbl
for operation and maintenance, for a total cost of $36/bbl. Thus, synfuel exploitation
could be costly. Raising CAFE standards is more effective (Section 15.2).
16.7.2 Gasohol
In 1978 Congress removed the 4 c/gal excise tax for gasohol, a fuel mix that is 10%
ethanol and 90% gasoline. Iowa followed with the removal of its 6.5 c/gal state
excise tax. The US subsidy for ethanol is
($0.04/gal)(42 gal/bbl)(100% gasohol/10% ethanol) = $17/bbl ethanol. (16.72)
Two decades later gasohol subsidies continued to play a role in the political
landscape, as presidential candidates praised gasohol before the early Iowa caucus.
However, in 2001 the National Academy of Sciences concluded that the extra credits
for multifuel vehicles had not been worthwhile. The federal subsidy for ethanol
for direct use and for farm subsidies totaled $30 billion between 1996 and 2002.
However, ethanol became competitive when oil hit $40–50 a barrel, but it will have
an impact on agriculture. Ethanol for fuel will need cellulosic technology to have
a major impact.
P1: GFZ
16.8 Plutonium Economy

A liquid metal fast breeder reactor (LMFBR) produces more plutonium than the
235
U or 239 Pu fuel they consume. (Plutonium proliferation is discussed in Chapter 5;
nuclear safety and waste disposal are discussed in Chapter 7.) The breeder reactor
was an issue in the 1970s and it remains an issue in 2003 as Russia and Japan
continue to aspire toward the breeder, while France and other European nations
use plutonium as mixed oxide (MOX) fuels in their light water reactors (LWR).
There is sufficient uranium to supply breeder reactors for billions of years, enough
to qualify LMFBRs as a renewable energy source, but there are other issues that
must be considered, for example, the economic one. In this section we will try to
answer the question, “Under what conditions can breeders produce electricity at
the same cost as light water reactors?”
It turns out that commercialization of liquid metal fast breeder reactors is more
likely to occur at such a time when three parameters come into play. They are
(a) high demand for nuclear power; (b) a decrease in LMFBR capital cost; and (c)
diminished uranium supplies. Conversely, the economic competitiveness of the
breeder remains in doubt as long as nuclear demand remains low, LMFBR capital
cost stays expensive and uranium supplies remain plentiful.
16.8.1 LMFBR Capital Cost

Estimates of the breeder’s capital cost place it 50% above the LWRs, $3000/kWe as
compared to $2000/kWe . Thus, the extra cost for an LMFBR is perhaps $1000/kWe .
The extra annual cost for a 1-GWe LMFBR at 10%/year (higher risk) is
(0.1/year)(106 kWe /GWe )($1000/kWe ) = $100 million/year. (16.73)
A 1-GWe LMFBR at 70% load factor produces
(106 kWe )(8766 h/year)(0.7 load factor) = 6.1 billion kWh/year, (16.74)
for an added cost per kilowatthour for LMFBR electricity of
($100 million/year)/(6.1 billion kWh) = 2 c/kWh. (16.75)
Extra costs for reprocessing and fuel fabrication should be added to give a total
extra cost of 3-4 c/kWh above the cost of electricity for the LWR. A LMFBR con-
sumes ordinary uranium, making its uranium ore essentially free, while an LWRs
consumes 5000–6000 tons of U3 O8 over its 30-year lifetime, or 200 ton/year. The
extra charges for just the LMFBR capital cost could be accommodated if uranium
for LWRs rose to a cost of
($100 million/year)(1/200 ton/year)(1 ton/900 kg) = $600/kg, (16.76)
which is far above the typical $20–30/kg uranium price. Hence, economic compet-
itiveness of the LMFBR is remote. If the price of uranium were to rise to $100/kg
(much less than the $600/kg transition price), there are alternative sources of
P1: GFZ
16.8. Plutonium Economy 423
Figure 16.7. The Canavan model for breeder commercialization. The commercialization
date of the LMFBR is determined by following the lines connecting the three graphs. See
text and Table 16.1 for details.
uranium, such as Chattanooga shale and, perhaps, the 4 billion tons of uranium in
seawater at 2 parts per billion.
16.8.2 The Canavan Model

Gregory Canavan’s (Los Alamos National Laboratory) model of LMFBR commer-
cialization combines three key parameters:
r Economics. The transitional price of uranium; it is determined by figuring under
what circumstances the cost of electricity from an LMFBR and the cost of an LWR
become equal.
r Geology. The amount of uranium available for the transition price; this is the size of
the available uranium resource.
r Energy Planning. The commercialization date for the LMFBR; this is determined
by availability of uranium, combined with energy demands placed on a reactor
and the rate at which society’s demand for nuclear energy grows.
The parameters used for the results of Fig. 16.7 come from the late 1970s when
the breeder debate raged. This happened during a time of high nuclear growth:
P1: GFZ
Table 16.1. LMFBR/LWR methodology. Calculations were

carried out in constant dollars and under the assumption of a
healthy nuclear industry in 1979. The 14 base-case parameters
are given below, along with the definition of 7 derived
parameters. LMFBR: Liquid metal fast breeder reactor; LWR:
light water reactors.
P1 = 5 GWe /yr, the future growth rate for nuclear power.
P2 = 5 million tons of U3 O8 at a price (2 × forward cost) of $100/lb.
P3 = $1000/kWe , the capital cost of an LWR reactor.
P4 = 200 tons U3 O8 /GWe -yr as fuel for LWR (load factor of 0.75).
P5 = $1500/kWe , the capital cost of an LMFBR reactor.
P6 = 0.60, the LWR load factor.
P7 = 0.60, the LMFBR load factor.
P8 = 0.4 c/kWh, the LWR fuel cycle cost (not U3 O8 ).
P9 = 0.5 c/kWh, the LMFBR fuel cycle cost.
P10 = 10%/yr, annual capital charge, all but fuel, no inflation.
P11 = 2 c/kWh, the cost of distribution, delivery, and losses.
P12 = 100 GWe , the size of the “first era” of nuclear power.
P13 = 1990, the first year of the “second era” of nuclear power.
P14 = 30 yr, the lifetime of a 1-GWe reactor.
P15 = transition price of U3 O8 ($/lb) at transition price P16 .
P16 = transition cost of electricity when LWR = LMFBR ( c/kWh)
P17 = tons of U3 O8 available at transition price P15 .
P18 = tons of U3 O8 used in the first nuclear generation.
P19 = tons of U3 O8 available for the second nuclear generation.
P20 = lifetime (yr) of the second nuclear generation.
P21 = commercialization date of the LMFBR.
Economics (solve for P15 and P16 ):
P16 = P8 + P11 + P10 P3 (1.11 × 10−8 )/P6 + P4 P15 (3.04 × 10−5 ).
P16 = P9 + P11 + P10 P5 (1.11 × 10−8 )/P7 .
Geology: P17 = (5 × 104 )P15 .
Energy planning (solve for P18 to P21 ):
P18 = P4 P12 P17 .
P19 = P17 − P18 = (0.5)P1 (P20 )2 .
P21 = P20 + P13 + P14 /2.
5–10 GWe /year. At that time it was estimated that there were 2.5–5 million tons
of uranium at $100/lb (including forward cost). Furthermore it was thought that
LMFBR capital cost would be 1.25–1.75 times the LWR cost.3 The graph labeled
Economics compares the LWR and LMFBR costs of electricity ( c/kWh) to the cost
of uranium ($/lb). The cost of electricity from LMFBRs is a horizontal line since
such reactors use little natural uranium. Conversely, electricity cost for LWRs show
a linear increase with the price of uranium since they consume much more natural
uranium. The point where these curves intersect is the transition price of uranium—
the price at which the LMFBR is able to compete economically with the LWR. The
3
Uranium reserves in 2001: <$30/lb (0.3 Mtons), <$50/lb (0.9 Mtons), <$100/lb (1.4 Mtons).
[Annual Energy Review, EIA., 2003]
P1: GFZ
Problems 425
graph labeled Geology displays a simplified version of the amount of uranium avail-
able versus uranium price. The upper curve is the larger potential resource and the
lower curve is the smaller probable resource. Extending the transition price of ura-
nium upward (from the Economics graph) to the geology graph gives the amount
of uranium available at the transition price. The last graph, labeled Energy Planning,
indicates the time-dependence of uranium consumption after initiation of the sec-
ond generation of nuclear power. Constant growth in GWe /year gives parabolic
uranium consumption proportional to t 2 . Extending the transitional uranium con-
sumption to the relevant growth curve gives the transition date for LMFBR com-
petitiveness. By this model the LMFBR commercialization date is 15 years prior
to the transitional year (denoted with the letter C), since one-half of a transitional
LWR-life is more expensive than the LMFBR.
These late-1970s projections do not reflect today’s halted nuclear power growth, in-
creased reprocessing costs, and cheaper uranium costs.4 The upper curve in Fig. 16.7 tells
the story, with (1) nuclear growth of 5 GWe /year, (2) optimistic uranium resources
of 5 Mton at $100/lb, and (3) higher LMFBR capital cost, 1.5 times the LWR. This re-
sult postpones LMFBR commercialization to 2118. The LMFBR commercialization
date can be brought forward to 2024 by (1) raising nuclear growth to 10 GWe /year,
(2) using a more pessimistic estimate of uranium resources of 2.5 Mtons at $100/lb,
and (3) lowering LMFBR capital cost to 1.25 times the LWR.
Problems
16.1 Present value. Compare the value of $1000 now at 8.1%/year interest versus
$1100 in 1.2 years.
16.2 Continuous versus annual interest. A $1000 bond matures to $2000 in 6
years. What is the interest rate with continuous, quarterly, and annual com-
pounding?
16.3 Tripling time. What is λTt for tripling time Tt and λ growth rate?
16.4 Fixed balance payments. The balance on a loan of Co is reduced linearly in
time, hence the borrowers payments are reduced over tine. The loan is paid
back in T years at interest i on the unpaid balance. Show that the average
annual payment is P = Co /T + Co i/2 and compare this result with a fixed
payment 30-year, $100,000 loan at 6% continuous interest.
16.5 Current versus constant dollars. The text uses a discount rate of about
3–5%/year, which is less than long-term interest rates. How can you jus-
tify this? What is the present value of saving $1000/year over 5 years and
30 years?
16.6 Elasticity. What is demand elasticity for gasoline if doubling gasoline prices
(only) mothballs 10% of 200 million US vehicles? What is the cross-elasticity if
4
Capital cost [nuclear ($2/We ), coal ($1.3/We ), natural gas ($0.5/We )]. Cost at busbar
[nuclear (6.7 c/kWh), coal (4.2 c/kWh), natural gas (4.1 c/kWh)]. Nuclear operation and
maintenance (1.5 c/kWh), capacity factor (85%), lifetime (40 yr) (Deutch and Moniz, 2003).
P1: GFZ
natural gas cars increase by 500% with a doubled gasoline price? How much
energy would be saved if natural gas cars (40 mpg) were to replace gasoline
cars (27.5 mpg), traveling 12,000 miles/year?
16.7 Gasoline prices. What was the annual average inflation rate between 1973
and 1999 if gasoline cost 30 c/gallon before the 1973 embargo and this cor-
responds (EIA) to $1.10 in 1999 dollars? The second oil shock of 1979 raised
gas to $2.05/gallon in 1999 dollars. What is this cost in 1979 dollars?
16.8 Payback periods. A $1000 retrofit insulation saves $200/year in fuel costs.
What is the payback period in current and constant dollars, without interest
and with interest of 7%/year and inflation of 3%/year? What is the payback
period using discounted benefits? What is the discounted benefit over 50–100
years?
16.9 Declined efficiency. (a) What is the payback period using discounted benefits
for problem 16.8, assuming insulation quality declines exponentially with age
as e−t/T with T of 100 years? (b) Redo but with a linear decline of t/T.
16.10 Cost of conserved energy. A 16-ft2 hot water heater with R5 insulation is
surrounded with an R7 blanket that costs $75. The water is kept at 120◦ F with
an average outside temperature of 60◦ F. What are the energy savings/year
and what is the cost of a conserved therm of natural gas (105 Btu)? How does
this cost compare to the price of natural gas in your area?
16.11 Minimum life cycle cost. Derive energy consumption at the minimum life-
cycle cost (Eq. 16.51).
16.12 Tax credits in FY 2001. What are the costs of conserved energy for the follow-
ing Clinton administration proposal in 2000 for tax credits using reasonable
scenarios and current costs. (a) A 15% tax credit for photovoltaic home units
(Section 13.1). (b) A 15% tax-credit for solar hot water heaters (Section 16.5).
(c) A $1000 tax credit for new homes that use 30% less energy than the Inter-
national Energy Conservation Code (use California standards, Section 11.4).
(d) A tax credit of $750–3000 for hybrid vehicles (Section 15.4). (e) A tax credit
of 1.5 c/kWh for electricity from windmills (Section 13.6).
16.13 Carbon tax. A US government report says that a $30/ton carbon “would
roughly stabilize carbon emissions at their 2000 levels and would generate
about $40 billion annually.” Show that $40 billion/year is reasonable? If we
assume that carbon use grows at 1–2%/year, what value of elasticity is needed
to “stabilize” carbon emissions? How much money would be raised with the
carbon tax?
16.14 Hubbert with economics. (a) What is the asymptotic long-range production
of oil using the parameters in Section 16.6 and e d = 0.3. (b) Use a Runge–Kutta
routine to reproduce some of the curves in Fig. 16.5
16.15 Indirect cost of imported oil. Using today’s web data, what is the indirect
cost of imported oil from OPEC with e s = 0.5−1? Begin with www.eia.gov.
16.16 Excise tax removal. When oil prices rose in 2000, Congress considered re-
moving the 4.3 c/gal excise tax. How much would this save a typical US
auto owner who drives 12,000 mi/year? How much tax revenue would be
lost? If e d = −0.1, how much more gasoline would be needed?
P1: GFZ
Bibliography 427
16.17 Buy a power plant. Deregulation of the electrical power industry convinced
utilities to sell power plants and buy others, presumably for depreciation
advantages. Stone and Webster stated that 145 GWe would probably be sold
at an average of $0.32/We between 1997 and 2001. Did this represent a good
or bad investment for 25% of US capacity?
16.18 Thermal recycle. A 1-GWe reactor burns 1000 kg/year 235 U and it produces
200 kg/year Pu in spent fuel, with one-third of the 100-ton core replaced
every 1.5 years. Reprocessing spent fuel and fuel fabrication for Pu MOX
fuel costs $2000/kg of heavy metal. However, Pu is only 75% as effective
as 235 U as a fuel. LWRs use only one-third MOX cores, but with redesign it
might be possible for the LWR to use full-MOX cores. A 1-GWe reactor needs
5000 tons of natural uranium over a 30-year lifetime. What is the break-even
cost of natural uranium for this process to be cost effective? U fuel costs
$100/kg SWU and $30/kg U. How many GWe year of MOX are in 70,000
tons of spent fuel at Yucca Mt with 1% Pu?
16.19 Breeder capital cost. A 1-GWe LMFBR saves 5000 tons of natural uranium
which a 1-GWe LWR consumes over a 30-year lifetime. If the extra cost of a
breeder is $500–1000/kWe (1980), what is the break-even cost of uranium to
build the breeder, including reprocessing and fuel fabrication at a total cost
of $2000/kg of heavy metal.
16.20 Free trade. David Ricardo (1772–1823) was the first economist to show that
trade barriers between nations prevent nations from specializing in their
best industries, lowering gross national products of all nations. (a) Develop
a simple mathematical model that takes into accounts tariffs between na-
tions. Consider the historic case example in Ricardo’s time, that of France,
which produces grain more efficiently than England, and England, which
made manufactured products more cheaply than France. Remove tariffs to
fractionally shift French labor toward agriculture and English labor towards
manufacturing. Using guesses for parameters, how much did French and En-
glish productivity increase? (b) What factors are ignored in your model and
by what means does the World Trade Organization attempt to overcome these
factors, especially in a world of industrialized and developing countries?
Bibliography
Brown M.A., M.D. Levine, J.P. Romm, et al. (1998). Engineering-economic studies of energy
technologies to reduce greenhouse gas emissions: Opportunities and challenges, Ann. Rev.
Buchholz, T. (1989). New Ideas from Dead Economists, Penguin, New York.
Daneke, G. (Ed.) (1982). Energy, Economics and the Environment, Lexington Books, Lexington,
MA.
Dargay, J. and D. Gatoly (1995). Response of world energy and oil demand to income growth
and changes in oil prices, Ann. Rev. Energy Environ. 20, 145–178.
Deutch, J. and E. Moniz (Eds.) (2003). Nuclear Power: An Interdisciplinary MIT Study, MIT
Press, Cambridge, MA,
P1: GFZ
Joskow, P. and D. Marron (1993). What Does Utility-Subsidized Energy Efficiency Really
Cost? Science 260, 281–370.
LeBel, P. (1982). Energy Economics and Policy, Johns Hopkins Univ. Press, Baltimore, MD.
Levine, M., et al. (1995). Energy efficiency policy and market failures, Ann. Rev. Energy
Environ. 20, 535–555.
Levine, M., et al. (1985). Economics of efficiency improvements in residential appli-
ances and space conditioning equipment, in Energy sources: Conservation and Renewables,
D.Hafemeister, et al. (Eds.), AIP Press, NY.
National Research Council (1986). Electricity in Economic Growth, National Academy Press,
Washington, DC.
——— (1993). Issues in Risk Assessment, National Academy Press, Washington, DC.
Pirog, R. and S. Stamos (1987). Energy Economics: Theory and Policy, Prentice Hall, Englewood
Cliffs, NJ.
Rosenfeld, A. (1999). The art of energy efficiency: Protecting environment with better tech-
nology, Ann. Rev. Energy Environ. 24, 33–82.
Samuelson, P. (1967). Economics: An Introductory Analysis, McGraw-Hill, New York.
Stobaugh R. and D. Yergin (Eds.) (1979). Energy Future: Report of the Energy Project at the
Harvard Business School, Random House, New York.
US Energy Information Administration (2003). Annual Energy Review, EIA, Washington, DC.
White House (2003). Economic Report of the President, Washington, DC.
Wilson, R. and E. Crouch (2001). Risk-Benefit Analysis, Harvard Univ. Press, Cambridge, MA.
P1: GFZ
SVNY342-Hafemeister February 9, 2007 18:13
Appendices
A. Nuclear Arms Chronology............................................................... 431
B. Energy/Environment Chronology ..................................................... 446
C. Units............................................................................................. 454
D. Websites........................................................................................ 460
E. Symbols......................................................................................... 464
F. Glossary......................................................................................... 469
G. Index ............................................................................................ 483
429
P1: GFZ
A. Nuclear Arms Chronology
1919
——June: Soon after World War I, Ernest Rutherford creates nuclei (4 He + 14 N ⇒
1
H + 17 O). “Talk softly please. I have been engaged in experiments, which suggest
that the atom can be artificially disintegrated. If it is true, it is of far greater
importance than a war.”
1920
——Rutherford speculates on existence of the neutron at the Royal Society (Lon-
don).
1931
——November: Harold Urey discovers deuterium in hydrogen’s optical spectra.
1932
——February: James Chadwick discovers the neutron. (4 He + 9 Be ⇒ 34 He + 1 n).
1933
——January 28: Adolph Hitler becomes Chancellor of Germany.
——March 4: Franklin Roosevelt becomes President of the United States.
——April: Eleven scientists who had won the Nobel Prize or would win it later
were fired from German Universities because they were Jewish. Johannes Stark
and Philip Lenard’s “Aryan physics” is not accepted by physicists.
——October: Leo Szilard recollects, “It occurred to me in October 1933, that a chain
reaction might be set up if an element could be found that would emit two
neutrons when it swallowed one neutron.” This idea becomes a classified British
patent in 1935 before uranium fission is discovered.
1934
——Artificial radioactivity discovered by Marie Curie and Frederic Joliet (alpha
particles) and by Enrico Fermi (neutrons).
1938
——December: Fermi receives Nobel prize for physics for discovery of transuranic
elements (actually fission of uranium) and departs for the “new world” of the
United States.
431
P1: GFZ
432 A. Nuclear Arms Chronology
——December 22: Otto Hahn, Fritz Strassman, Lise Meitner and Otto Frisch con-
clude that identification of barium implies that uranium nucleus was fissioned
by neutrons.
1939
——January to May: Many experiments on uranium fission.
——April 29: Conference in Berlin to consider a German bomb and nuclear reactor.
——August 2: Szilard and Edward Teller obtain a letter from Einstein on the possi-
bility of uranium weapons, which Roosevelt receives from Alexander Sachs on
October 11, 1939.
——September 1: Hitler invades Poland.
1940
——June 3: Paul Hartek fails to produce neutron multiplication in Hamburg reactor
(185 kg natural UO2 , 15 tons CO2 ice).
1941
——January: Experiments at the Berlin “virus house,” using a 300-kg uranium
reactor with impure graphite, cause incorrect rejection of graphite as a moderator
in favor of heavy water from Vermork, Norway.
——July: British “Maud” Committee reports that a weapon could be made with
10 kg 235 U; US National Academy Sciences endorses bomb program.
——August: Hamburg group begins construction of ultracentrifuges to obtain 235 U.
Centrifuges explode in April 1942, but they attained 7% enrichment by March,
1943.
——October: Neils Bohr and Werner Heisenberg indirectly discuss nuclear
weapons, described in Michael Frayen’s play Copenhagen.
——December 6: Roosevelt directs substantial financial and technical resources to
construct uranium bombs.
——December 7: Japan attacks Pearl Harbor.
1942
——May: Germans observe first neutron multiplication (13%) in Leipzig using
570-kg U and 140-kg heavy ice.
——December 2: First nuclear chain reaction at Chicago’s Stagg Field by Fermi.
1943
——February 28: Vermork heavy water factory destroyed by Allies after a failed
attack on November 19, 1942.
——March 15: J. Robert Oppenheimer moves to Los Alamos, New Mexico.
——March: Glen Seaborg suggests Pu weapons might be jeopardized by isotope
240
Pu if it is a spontaneous neutron emitter.
——Resumption of Soviet nuclear experiments.
1944
——August: Bohr presents memorandum on international control of nuclear
weapons to British Prime Minister Winston Churchill and Roosevelt.
——November: First batch of spent fuel obtained from Hanford reactors.
P1: GFZ
A. Nuclear Arms Chronology 433
——November: Samuel Goudsmit’s Alsos mission obtains documents in Stras-

bourg which imply that German weapons program had made little progress.
1945
——January: First Pu reprocessing production at Hanford.
——January 20: First 235 U separated at the K25 gaseous diffusion plant, Oak Ridge,
Tennessee.
——April 25: UN Charter signed by 50 nations in San Francisco.
——May 4: End of war in Europe.
——June 11: The Franck Report on the demonstration of the bomb and its interna-
tional control sent to US Secretary of War.
——July 16: US explodes first atomic bomb, Trinity, at Alamogordo, NM with
electronics shielded to avoided EMP pulse.
——August 6, 9: Atomic bombs destroy Hiroshima (Thin Man uranium, 13 kton,
3 m by 0.7 m diameter) and Nagasaki (Fat Man plutonium, 22 kton, 3.3 m by
1.5 m). Each weapon kills 70,000 prompt deaths and 140,000 within a year.
——August 15: End of war in the Pacific.
1946
——June 14: Bernard Baruch presents to the UN the Dean Acheson and David
Lilienthal plan to internationalize the atom. “We are here to make a choice be-
tween the quick and the dead.”
——June 30: First subsurface detonation by US at Bikini Atoll.
——July: Demonstrations in Times Square, New York City, against nuclear
testing.
——December 31: Atomic Energy Commission takes over nuclear weapons pro-
gram from US Army.
1948
——April, May: US atomic tests, Eniwetok Atoll in the Marshall Islands.
——November 4, 1948: UN General Assembly adopts US plan for international
atomic control; USSR is opposed.
1949
——April 4: NATO established.
——August 29: First Soviet atomic detonation, in the Ustyurt Desert.
——October 30: AEC General Advisory Committee gives approval to build more
powerful U bombs rather than the H bomb.
1950
——January 27: Klaus Fuchs confesses he transmitted atomic secrets to the Soviets.
——January 31: US President Harry Truman announces the decision to proceed
with the H bomb.
——March: Worldwide peace offensive to “ban the bomb” (the Stockholm Appeal)
signed by more than 500 million people.
——June 25: North Korean army crosses 38th parallel, starting a war that lasted
until July 1953.
P1: GFZ
1952
——January: UN Security Council establishes Disarmament Committee.
——June: Lawrence Livermore Laboratory established.
——October 3: First British nuclear test, Monte Bello Islands, Australia.
——October 31: US explodes first fusion device Mike of 10 Mtons at Eniwetok
(liquid deuterium, not deliverable).
1953
——March: US above-ground tests begin in Nevada.
——May: Utah officials asks AEC to stop Nevada tests because of fallout on
St. George, Utah.
——August 12: First Soviet fusion device exploded on a tower in Siberia (probably
not deliverable).
——Fall: India and Australia propose a total ban on nuclear weapons to the UN.
——December 8: “Atoms for Peace” speech by US President Dwight Eisenhower
at UN.
1954
——January 21: First nuclear-powered submarine, USS Nautilus, launched.
——March: “Bravo” H bomb test affects Marshall Islanders with fallout.
——April: British Parliament petitions Churchill, Eisenhower, and Malenkov to
control nuclear weapons.
——April 12 to May 6: AEC hearings deny Oppenheimer his access to nuclear
secrets.
——August 30: Atomic Energy Act of 1954 emphasizes peaceful uses of atomic
energy.
1955
——January–December: Reports of increasing global nuclear fallout.
——May 6: West Germany joins NATO.
——May 14: Warsaw Pact Organization established.
——August 8–20: First International Conference on Peaceful Uses of Atomic Energy
in Geneva.
——November 23: First deliverable high-yield Soviet H bomb.
1956
——US National Academy of Sciences panel finds genetic effects from fallout of nu-
clear tests to be slight compared with effects from natural radiation background;
Adlai Stevenson makes fallout an issue in US presidential campaign.
1957
——May 15: First British H bomb exploded at Christmas Island in the Pacific Ocean.
——July 6–11: First Pugwash Conference advocates test ban; Soviet scientists at-
tend.
——August: AEC inaugurates Plowshare Program for peaceful nuclear explosions.
——September 19: Rainier, the first underground test at 1.7 ktons at Nevada Test
Site.
P1: GFZ
——October 1: International Atomic Energy Agency inaugurated in Vienna.

——October 4: First satellite Sputnik I put into orbit by the USSR.
——December 2: Shippingport electric power reactor reaches full power of 60 MW.
1958
——January 1: European Atomic Energy Community (Euratom) established.
——January 31: Explorer I, first US satellite, put into orbit.
——November 1958–September 1961; US, UK, USSR observe informal moratorium
on nuclear tests.
1959
——September 2: Vela Uniform Seismic Project established by US Department of
Defense.
——November 24: US/USSR sign a memorandum of cooperation for utilization of
nuclear energy.
1960
——February 13: First nuclear test by France in the Sahara Desert.
——May 1: U-2 spy plane shot down over USSR.
——July 26: USSR suggests 3 on-site inspections/year as part of a test ban.
——November 15: First Polaris missile launched from a US submarine.
1961
——January 17: Eisenhower’s farewell address discourses on “the military-
industrial complex.”
——February 1: US launches Minuteman-I missile.
——March 31: Pravda suggests the use of H bombs to obtain fresh water from
Soviet glaciers.
——April 12: Yuri Gagarin becomes the first cosmonaut in orbit.
——May 29: US and UK agree to draft Comprehensive Test Ban (CTB) Treaty with
12 on-site inspections/year.
——August: Installation of first US seismic stations in a network of 60 countries.
——September 1: USSR resumes nuclear tests, including a 58-Mton explosion on
Oct. 31.
——September 15: US resumes nuclear tests.
1962
——February 20: John Glenn becomes first US astronaut in orbit.
——July 6: Sedan excavation experiment with Plowshare peaceful nuclear explo-
sion; 10 million tons of earth displaced.
——July 8: Electromagnetic pulse from “Fishbowl” test (1.4 Mtons, 400 km
above Johnston Island) turns off 300 streetlights on Oahu, Hawaii (1200 km
away).
——September 3–7: Tenth Pugwash Conference proposes “black box” monitoring
for CTB treaty verification.
——October: US reduces requirement to 8–10 inspections/year for a CTBT.
——October 23 to November 20: Cuban missile crisis and blockade.
P1: GFZ
——December 19: Communist Party Secretary Nikita Khrushchev accepts 2–

3 inspections/year and 3 unmanned seismographic stations (black boxes) in
USSR.
1963
——January: Secret test ban talks between US and USSR.
——February 19: US accepts seven inspections/year, provided any mysterious
event may be challenged.
——April: Khrushchev withdraws offer of three inspections/year.
——June 20: US and USSR sign “hot line” agreement.
——August 5: US/USSR/UK sign Limited Test Ban Treaty, banning nuclear explo-
sions in atmosphere, space, and underwater.
1964
——August 2: Gulf of Tonkin resolution sends US troops to Vietnam under Presi-
dent Lyndon Johnson.
——October 16: China explodes first nuclear bomb.
——October 22: Salmon, a US cavity-explosion, muffles seismic signal in a salt
dome, near Hattiesberg Mississippi.
1965
——June: Large Seismic Array (LASA, Montana) starts detecting nuclear explo-
sions.
1966
——January 17: B-52 bomber crashes near Palomares, Spain, with 4 unarmed H
bombs, two by sea and two by land.
——September 24: First French H bomb, in Tuamato Archipelago, Polynesia.
1967
——January 27: Outer Space and Celestial Bodies Treaty bans nuclear weapons in
orbit and space.
——February 14: Treaty for Prohibition of Nuclear Weapons in Latin America
signed in Mexico City (Tlatelolco); all nations must ratify for the treaty to en-
ter into force.
——June 17: First Chinese H bomb exploded in the atmosphere.
——June 23–25: President Johnson and Soviet Premier Alexi Kosygin begin SALT
and ABM treaty talks in Glassboro, NJ.
——December 10: Gasbuggy Plowshare explosion in New Mexico to stimulate
natural gas production.
1968
——July 1: Nuclear Non-Proliferation Treaty (NPT) opened for signature.
——August: Soviet GALOSH ABM system deployed around Moscow.
1969
——July 20: US Apollo 11 lands on moon.
P1: GFZ
——November 3: UN Committee on Disarmament proposes global exchange of

seismic data to monitor a CTB treaty.
——November–December: Preliminary SALT talks in Helsinki.
1970
——January: Administration of President Richard Nixon cuts Plowshare budget.
——March 5: NPT enters into force with 50 nations in 1970 and 188 in 2006.
——November 30: Atlantic-Pacific Interoceanic Canal Commission rejects use of
nuclear explosives.
1971
——March 30: US deploys Poseidon SLBM.
——June 15: Three Russian nuclear explosions set off to increase oil productivity.
1972
——May 26: SALT I and ABM treaties signed by Nixon and Soviet First Secretary
Lenoid Brezhnev in Moscow.
——June: UN Conference votes 48 to 2 to halt nuclear testing; France/China op-
pose, US/UK abstain, USSR absent.
——October: US detects Soviet SS-18 flight test.
——November: SALT II negotiations begin.
1973
——May 30: Protocol establishes the ABM/SALT Standing Consultative Commis-
sion.
——June 22: World Court issues an injunction to France to refrain from testing on
Mururoa in Pacific Ocean.
——October 17, 1973, to March 17, 1974; OPEC embargoes oil to US/Netherlands.
1974
——May 18: India sets off a 10-kton device under Rajasthan Desert, expanding
nuclear club beyond WWII “big five.”
——November 24: US/USSR agree to limit strategic launchers (to 2400) and MIRV
launchers (to 1320).
1975
——January 19: AEC reorganized into NRC (regulatory) and ERDA (development,
became DOE in 1977 and NNSA in 1999).
——October 1: Grand Forks, North Dakota Safeguard ABM site is activated. It was
closed 4 months later because of its ineffectiveness against Soviet MIRVs at a cost
of $21 billion from 1968–78 (1998$).
1976
——January 29: France cancels its reprocessing plants for South Korea and Pakistan
(1978).
——October 28: US President Gerald Ford postpones reprocessing of spent fuel to
obtain Pu.
P1: GFZ
1977
——April 7: US President Jimmy Carter postpones indefinitely commercial repro-
cessing, halts Clinch River Breeder Reactor and calls for an International Fuel Cy-
cle Evaluation (INFCE) to make the nuclear cycle more resistant to proliferation.
——October 19, 1977, to February 26, 1980: 40 nations prepare INFCE report.
1978
——January 11: Nuclear Suppliers guidelines forwarded to IAEA.
——March 10: Nuclear Nonproliferation Act (NNPA) tightens nuclear export cri-
teria.
——April 4: Anwar Saddat of Egypt and Menahin Begin of Israel sign Camp David
Accord in Washington.
——July 13: Euratom ends US embargo of nuclear fuel to Europe (NNPA required
renegotiation) with a brief meeting.
1979
——April, 6: US cuts economic and military aid to Pakistan because it is making
weapons-grade uranium.
——June 8: First Trident with SLBMs launched.
——June 18: Brezhnev and Carter sign SALT II Treaty in Vienna.
——September 22: Vela satellite bhangmeter picks up a signal over the South
Atlantic, perhaps a 3 kton explosion.
——December 12: NATO agrees to negotiate INF Treaty, while building Pershing-II
and ground launched cruise missiles.
——December 26: USSR invasion of Afghanistan removes SALT II from US Senate
consideration.
1980
——July 15: Office of Science and Technology Policy concludes the light signals
recorded by a Vela satellite over the South Atlantic on September 22, 1979, were
probably not from a nuclear explosion. Others disagree.
——December: Tomahawk submarine-launched cruise missile flies 600 miles to sea
and returns to land with 5-m accuracy.
1981
——April 12–14: Space shuttle Columbia orbits Earth 36 times.
——June 7: Israel destroys Iraq’s Osirak reactor, an act that was condemned at the
time but appreciated after the Gulf War.
——November 13: Senate ratifies Protocol I of Tlatelolco Treaty; US cannot deploy
weapons in Puerto Rico, Virgin Islands, or Guantanamo Naval Base.
——November 30: US tables Zero-Zero INF proposal in Geneva.
1982
——June 14, Argentina surrenders to UK in the Falkland War
——June 29: Strategic Arms Reduction Talks (START) begin in Geneva.
——July 16: Ambassadors Paul Nitze and Yuli Kvitsinsky walk in the Geneva
woods, trying to cut INF weapons by 50%.
P1: GFZ
——1982: USSR places 300 SS-20 missiles on its western border.

——November: US President Ronald Reagan adopts dense-pack basing for MX
missiles.
1983
——March 23: Reagan calls for SDI “to make nuclear weapons impotent and ob-
solete.”
——April 6: Scowcroft Commission recommends Midgetman (road-mobile, single-
warhead) and 100 MX.
——Summer: Large phased-array radar construction discovered at Krasnoyarsk.
——November 23: Soviets leave Intermediate Nuclear Force (INF) negotiations as
Pershing-II and GLCM are deployed.
—— December 8: Soviets suspend START negotiations.
1984
——January: Strategic Defense Initiative Organization (SDIO) created.
——January 23: Reagan’s first “Soviet Noncompliance” report released to Congress
amid controversy.
——January 30: Soviets respond with an aide memoire to US State Department on
US noncompliance.
——June 10: Homing Overlay Experiment (HOE) intercepts midcourse missile in
limited test of kinetic kill vehicles.
1985
——January 8: US Sec. State George Schultz and Soviet Amb. Andrei Gromyko
agree to negotiate on START/INF/ ABM/SDI.
——February 20: Nitze criteria for SDI deployment in terms of reliability, surviv-
ability, and marginal cost effectiveness.
——October 6: Reagan encourages broad interpretation of ABM Treaty to allow
tests of other physical principles in space.
——October 14: Schultz states that a broad interpretation the ABM Treaty “is fully
justified” to allow SDI research testing.
1986
——July: Natural Resources Defense Council sets up seismographs in the Soviet
Union.
——September: First B-1 bomber squadron becomes operational.
——October 7: Mordechai Vanuu publishes photos of the Israeli Dimone nuclear
weapons facility in the London Times. He is drugged and taken to Israel for
solitary confinement until April 22, 2004.
——October 11–12: Reagan and Soviet leader Mikhael Gorbachev agree to cut
strategic weapons by 50% in START.
——November 28: The 131st B-52 bomber with air launched cruise missiles is
deployed, exceeding the 1320 warhead limit on MIRVed launchers, negating
SALT II. US is first country to publicly renounce an arms control agreement.
——December: First 10 Peacekeeper missiles become operational in Minuteman-III
silos in Wyoming.
P1: GFZ
1987
——February: Soviets end unilateral nuclear testing moratorium after 18 months.
——April 7: Missile Control Technology Regime signed to slow missile prolifera-
tion.
——April 23: APS Directed Energy Weapons Study (DEWS) Group “estimates that
even in the best of circumstances, a decade or more of intensive research would
be required [for an informed decision on DEWS].”
——August: First of six South African gun-type weapon is qualified.
——September 15: Schultz and Soviet Foreign Minister Eduard Shevardnadze sign
Nuclear Risk Reduction Accord. Soviets present a list of permitted criteria for
testing under the ABM treaty.
——December 8: Reagan and Gorbachev sign Intermediate Nuclear Force Treaty.
1988
——May 27: US Senate ratifies INF Treaty (93–5), the first ratification of an arms
control treaty since 1972.
——June 28: Office of Technology Assessment (OTA) testifies it disagrees with the
charge of “likely violation of TTBT.”
——August–September: Joint Verification Experiment on nuclear tests carried out
by US/USSR.
1989
——November 9: Fall of Berlin Wall.
1990
——October 24: Last Soviet nuclear test, as Gorbachev calls for a moratorium.
——November 19: NATO and Warsaw Pact sign Conventional Armed Forces in
Europe (CFE) Treaty, limiting tanks/planes to same levels, a 60% reduction for
WTO and no reduction for NATO. Enters into force in 1992, ending the Cold War.
1991
——April 3: UN Resolution 687 requires destruction of Iraq’s nuclear/bio/chem
weapons.
——July 10: South Africa signs NPT and later destroys its six nuclear weapons.
——July 31: US and USSR sign START I, each cutting from 12,00 to 7000 actual
warheads.
——September 27/October 5: US President George H.W. Bush and Soviet leader
Gorbachev reciprocal-unilateral measures place missiles to be cut by START I off
alert, and remove surface-ship and ground-based tactical weapons.
——November 27: Sam Nunn–Richard Lugar Act passes, giving Russia first $400
million to destroy weapons.
—— December 1: Ukraine votes to become independent of USSR, followed by 13
other Soviet Republics and Russia.
1992
——March: US ends military Pu reprocessing, HEU production stopped in 1964.
——March 9: China joins the NPT regime.
P1: GFZ
——March 24: Open Skies Treaty signed, allowing monitoring by aircraft over-
flights of NATO and Warsaw Pact nations.
——May 23: START’s Lisbon Protocol signed by US/Russia/Ukraine/Belarus/
Kazakhstan, with all agreeing to join NPT.
——August 3: France joins NPT regime.
——September 23: Last US nuclear test.
——October 1: US Senate ratifies START I with the Biden condition calling for “the
adoption of cooperative measures to enhance confidence in reciprocal declara-
tions [for warheads and fissile materials].”
1993
——January 3: US/Russia sign START II, cutting deployed strategic warheads from
7000 to 3500.
——February 18: US/Russia agree on sale of 500 tons of HEU, blended to LEU
reactor fuel.
——September: US/Russia agree to start Materials Protection Control and Ac-
counting program with lab-to-lab assistance.
——September 27: President Bill Clinton proposes Fissile Material Cut-off Treaty
banning production of weapons Pu/HEU .
——December: UN adopts resolution calling for a Fissile Material Cut-off Treaty.
1994
——April 30: Nuclear Proliferation Prevention Act (Glenn-Pell Act) provides for
sanctions on nuclear testing.
——May 30: US/Russia detarget strategic missiles, but they can be quickly repro-
grammed.
——June 23: US Vice President Gore and Russian Prime Minister Viktor Cher-
nomyrdin agree to close last 3 Russian Pu-production reactors by 2000 with
some US aid to establish another power source.
——September 20: Clinton and Russian President Boris Yeltsin agree to exchange
data on warheads and Pu/HEU.
——October 1: Russia stops plutonium production, except at 3 reactors where waste
heat is used to heat cities.
——October 23: US/N. Korea sign “Agreed Framework” to eliminate N. Korean
nuclear program and get S. Korean reactors.
1995
——March 1: Clinton announces permanent removal of 100 tons of Pu/HEU from
US stockpile.
——May 10: Clinton and Yeltsin agree to declarations/inspections on excess
Pu/HEU from weapons.
——May 12: NPT extended indefinitely after the five nuclear weapon states promise
to join a CTBT.
——August 11: Clinton announces US will support a zero-yield CTB.
——December 19: Wassenaar Arrangement on export controls replacing Cold War
controls.
P1: GFZ
1996
——January 26: Senate ratifies START II by 87–4, cutting to 3500 total warheads
each and banning land-based MIRVs.
——January 27: France conducts last test and supports a zero-yield CTB.
——March 25: US joins South Pacific Nuclear-Free Zone.
——April 11: Pelindaba Treaty for Nuclear-Free Africa signed by 43 African states.
——September 10: UN opens CTB Treaty for signature, banning nuclear tests ev-
erywhere, for all time, at any yield. Only India, Bhutan, and Libya are opposed.
——September 17: US/Russia/IAEA agree to trilateral inspections on excess SNM.
——September 24: All five nuclear weapon states sign the CTBT at the UN.
1997
——April 29: Chemical Weapons Convention enters into force.
——March 21: Clinton and Yeltsin in Helsinki call for START III reductions to 2000–
2500 deployed, strategic warheads and “measures relating to the transparency”
for inventories and destruction of strategic warheads.
——May 27: Yeltsin signs NATO-Russian Founding Act.
1998
——May: India and Pakistan conduct multiple underground nuclear tests.
——August 31: North Korea launches a Dong-1 missile over Japan.
——September: US/Russia establish the Nuclear Cities Initiative.
——September 7: General Lebed reveals that 100 “suitcase-sized” nuclear weapons
may been “lost” by Russia.
——September 22: US and Russia sign Nuclear Cities Agreement to help 10 Russian
closed weapon cities.
1999
——October 13: US Senate rejects CTBT (51–48) in spite of its 150 signatory nations.
2000
——April: Duma ratifies CTBT (April 21) and START II (April 14), contingent on
ABM demarcation to constrain defenses.
——September 1: Gore/Chernomyrdin agree to dispose of 34 t Pu each, as MOX
fuel or buried in geological repository.
2001
——May 1: President George W. Bush states he will withdraw US from ABM Treaty
to test/deploy unclear ABM options.
——May 31: INF continuous Perimeter and Portal Monitoring ends as INF 13-year
period ends.
——September 11: Osama bin Laden’s al Qaeda uses US commercial planes to kill
3000 in destroying the World Trade Center and damaging the Pentagon. The US
joins Afghan resistance forces to topple Taliban government, producing a new
leader, Hamid Karzai. The 9-11 event creates a major paradigm shift to focus on
terrorism over other matters.
P1: GFZ
——September-October: Letter-borne anthrax contamination in US postal system

and buildings.
——December: G. W. Bush administration funds CTBT research, but not that which
aids CTBT entry-into-force.
——December 4: US and Russia complete START I reductions to 6,000 accountable
warheads.
——December 7: US ends discussion on verification of Biological Weapons Con-
vention.
——December 13: G.W. Bush formally starts process to withdraw from ABM Treaty
6 months hence.
——December 14: Navy Area Theater Ballistic Missile Defense system, covering
small areas, is cancelled.
2002
——January 4: DoD Ballistic Missile Defense Office elevated to Missile Defense
Agency.
——January 8: Nuclear Posture Review sets path to 1700–2200 deployed, opera-
tional US warheads by 2012, while retaining up to 5000 warheads on “active
and inactive reserve” that can be deployed in days to years. The “new triad”
combines offense, defense, conventional arms, and rapid information. Targeted
countries are Russia, China, Iran, Iraq, Libya, N. Korea, and Syria in possible
conflicts between Israel/Iraq, N. Korea/S. Korea, PRC/Taiwan. The NPR sug-
gests that new designs are needed to create a 5-kton earth-penetrating bomb to
attack underground bunkers. Congress shortens preparation time for test-sites
readiness for this weapon.
——January 23: Bush administration continues US/Russia Pu-MOX disposition
but cancels geological immobilization.
——January 29: Bush labels North Korea, Iran, and Iraq as an “axis of evil arming
to threaten the peace of the world.”
——March 14: India integrates the Agni-2 missile (2000 km, 1000 kg) into its armed
forces.
—— May 24: Presidents George W. Bush and Vladimir Putin sign the Strategic
Offensive Reductions Treaty, which reduces operational warheads to 1700–2200.
SORT allows response forces of some 2400 additional warheads, adds no new
verification measures, and expires on December 31, 2012, the same date the limit
of 1700–2200 is required. Russia retains 138 SS-18s and many SS-19s which would
have been removed under START II.
——October 16: US announces North Korea admits having a clandestine centrifuge
program to make HEU.
——November 4: Cuba accedes to the NPT, completing Western Hemisphere. UN
preparatory conference lobbies nuclear weapon states to agree to a negative
security assurance, so nonweapon states would not fear nuclear weapons.
—— November 27: UN inspections begin in Iraq, pursuant to UN resolution 1444
that required Iraq to disarm its weapons of mass destruction, or be in “material
P1: GFZ
breach.” In 1997, Iraq declared eight presidential palaces off limits, leading to an
end of inspection.
—— December 8: Iraq declares its lack of weapons of mass destruction as required
by UN resolution.
2003
——January 11: North Korea withdraws from NPT after IAEA inspectors are re-
moved (Dec. 27) and Governing Board “deplores in the strongest terms” (Jan. 6)
the reprocessing restart, claimed two weapons and beginning centrifuges.
——March 2: Department of Homeland Security established with a budget of
$30 billion within a year.
——March 6: Senate ratifies Strategic Offensive Reduction Treaty. Iraq-war anger
delays Russian Duma to May 14, 2003.
——March 19: US/UK invade Iraq for a 3-week war, without explicit support from
the UN, which wanted more time for more inspections. Unrest in Iraq continues
after the war, without evidence for weapons of mass destruction.
——March 31: US mildly sanctions Pakistan’s Kahn Research Institute for exporting
centrifuge equipment to N. Korea in trade for missile technology.
——April 22: Los Alamos builds first US Pu pit in 15 years, a beginning for a plant
to make 125 pits/year by 2018.
——April 25: North Korea tells US in Beijing that it has Nuclear weapons.
——June 6: IAEA Board of Governors charges Iran with a violation over clandestine
centrifuges discovered on February 20.
——August 28: North Korea informs China, Japan, Russia, South Korea, and the
US that it may test in the near future.
——December 14: Saddam Hussain captured.
2004
——February 4: Pakistan President Pervez Musharraf pardoned Abdul Qadeer
Khan for supplying centrifuge technology and equipment to Iran, Libya, and
North Korea.
——February 20: IAEA report stated that Libya had ordered 10,000 centrifuge tubes,
but was now cooperating.
——July 9: The Senate Intelligence Committee Report rebukes the October 2002 Na-
tional Intelligence Estimate that “Baghdad has chemical an biological weapons
by stating that this conclusion “overstated both what was known and what intel-
ligence analysis judged abut Iraq’s chemical and biological weapons holdings.”
The CIA did not invite DOE to test alleged centrifuge tubes. The administration’s
charge of Iraq’s purchase of uranium was incorrect.
——July 22: The 9/11 Commission Report states that “we have seen no evidence that
these or other earlier contacts ever developed into a collaborative operational
relationship [between Al Qaeda and Iraq].”
——September 30: Report on Iraq Weapons of Mass Destruction released by Charles
Duelfer (Chief Iraq WMD Inspector). It agreed with the interim results of David
Kay (“we were all wrong”). The report concluded that Iraq’s nuclear program
ended in 1991, that Iraq unilaterally destroyed its chemical weapons in 1991 and
P1: GFZ
that Iraq had abandoned its ambitious biological weapon plans. The report states
that Saddam retained his desire for WMD.
——October: CTBT ratified by 119 nations (not US, China) of 173 signatories (not
India, N. Korea, Pakistan).
——October 19: Brazil agrees to IAEA inspections of its centrifuges. This is hoped
to be a precedent for Iran.
——December 7: US intelligence community organized under a centralized Na-
tional Intelligence Director.
2005
——January 31: Elections held in Iraq with high Shiite and Kurd turn-out but
negligible Sunni participation.
——February: Iran nuclear program becomes evident as France/Germany/UK try
diplomacy and US exerts pressure.
——February 10: North Korea claims it has nuclear weapons and refuses to join
6-party talks.
——March 6: China questions US data on 8 or 9 DPRK nuclear weapons and sug-
gests 2-party, US/DPRK talks.
——July 29: State department absorbs its arms control bureau into the non-
proliferation bureau.
2006
——March 2: President Bush and Prime Minister Manmohan of India agree to
cooperate on peaceful nuclear power, but this requires amending the 1978 non-
proliferation law and obtaining unanimous consent from the Nuclear Suppliers
Group. India would place 14 of its 22 reactors under India-specific IAEA safe-
guards.
——April: Iran declares it has 164 centrifuges. US joins other France, Russia, and UK
in an offer to dissuade Iran from enriching uranium on its soil. Brazil continues
its work since 2004 on its enrichment plant.
——October 9: N. Korea detonates a nuclear explosion of less than 1 kt. Radiation
confirmed on Oct. 11.
P1: GFZ
B. Energy/Environment Chronology
1859
——Edwin Drake drills 21 m for 500 bbl of oil at Titusville, Pennsylvania, to begin
the petroleum era.
1942
——December 2: Enrico Fermi’s reactor goes critical at Stagg Field, University of
Chicago, Illinois to begin the nuclear era.
1945
——August 15: Office of Price Administration (OPA) lifts gasoline rationing.
1946
——May 6: Division of Oil and Gas established in Department of Interior.
——May 21: US President Harry Truman orders US Government to take possession
of coal mines during a strike.
——June 18: National Petroleum Council established.
1947
——January 1: Atomic Energy Commission begins operation.
——March 25: Coal-mine disaster kills 11 in Centralia, Illinois.
——June 16: Federal Power Commission authority extended to all natural gas
producers.
1952
——December 5: Severe air pollution (0.7 ppm SOx and particulates) kills 4000 in
London in 4 days.
1953
——August 7: Congress gives US government jurisdiction of ocean floors beyond
3-mile boundary.
——December 8: US President Dwight Eisenhower delivers “Atoms for Peace”
speech before the United Nations.
446
P1: GFZ
B. Energy/Environment Chronology 447
1954
——August 30: Atomic Energy Act of 1954 encourages peaceful use of nuclear
energy.
1957
——King Hubbert correctly predicts US petroleum production peak between 1966–
71, which happened in 1970 when lower forty-eight produced 9.1 Mbbl/day.
Hubbert uses a finite resource in differential equations, but does not use eco-
nomics.
1959
——March 10: Eisenhower limits oil imports to stimulate domestic production and
refining capacity.
1962
——October 11: Congress authorizes the president to impose mandatory oil import
quotas.
1963
——December 17: Clean Air Act provides assistance to states for air pollution
research.
1965
——October 2: Water Quality Act establishes the Water Control Administration.
——October 20: Solid Waste Disposal Act provides assistance for study, collection,
and disposal of solid wastes.
——November 9: First major power blackout covers northeast US.
1967
——November 21: Clean Air Act gives authority to Secretary of Health Education
and Welfare to set autoemission standards.
1969
——January 1: National Environmental Policy Act (NEPA) establishes framework
for Environmental Impact Statements and the Council of Environmental Quality
(CEQ).
——January–February: Major oil spill from offshore drilling near Santa Barbara,
CA.
——December 30: Oil depletion allowance reduced from 27.5% to 22%.
1970
——March 5: President Richard Nixon issues executive order requiring federal
agencies to evaluate their activities under the National Environmental Policy
Act.
——April 22: First “Earth Day” celebration.
——July 9: Nixon requests Congress to create Environmental Protection Agency
(EPA).
——October 23: Merchant Marine Act Amendment provides subsidies for oil and
liquified natural-gas tankers.
P1: GFZ
448 B. Energy/Environment Chronology
——December 24: Geothermal Steam Act authorizes leases for geothermal

steam.
——Clean Air Act sets national air quality and autoemission standards.
1971
——July 23: Supreme Court decision on siting of nuclear power plant at Calvert
Cliffs, Maryland, requires the Atomic Energy Commission to comply with the
National Environmental Policy Act.
1972
——EPA bans DDT.
——Clean Water Act sets pollution standards for water.
——US and Canada agree to clean up the Great Lakes, source of 95% of US fresh
water, used by 25 million persons.
1973
——June 29: White House Energy Policy Office created with former Governor John
Love of Colorado as first “energy czar.”
——October 17, 1973 to March 17, 1974: Organization of Petroleum Exporting
Countries (OPEC) embargoes US and the Netherlands because of their support
for Israel.
——November 7: Nixon creates Project Independence to end oil imports by
1980.
——November 27: Emergency Petroleum Allocation Act provides authority for oil
allocations.
——December 4: Presidential Federal Energy Office created with William Simon
as energy czar.
——December 15: Congress mandates daylight savings to save energy.
——December 28: Nixon signs Endangered Species Act.
——EPA begins phasing out leaded gasoline.
——EPA begins limits on factory pollution discharges.
1974
——June 22: Federal Energy Administration (FEA) authorized to order utilities and
industry to convert from oil/gas to coal.
——September 3: Congress authorizes funds for geothermal energy and solar heat-
ing demonstrations.
——October 5: Congress repeals daylight savings time in winter to save energy.
——October 11: Energy Reorganization Act abolishes the AEC to create the Energy
Research and Development Administration (ERDA, later DOE) and the Nuclear
Regulatory Commission (NRC).
——December 31: Congress requires ERDA to submit an annual comprehensive
energy plan.
——Arthur Rosenfeld establishes research on buildings and energy at Lawrence
Berkeley National Laboratory to begin the enhanced end-use efficiency energy era.
Rosenfeld was the last graduate student of Enrico Fermi, who began the nuclear
P1: GFZ
reactor era in 1942. The LBL work was originally based on the 1974–75 study by
the American Physical Society.
1975
——January 4: Congress establishes the 55-mph speed limit to save energy.
——March 17: Supreme Court rules that the states do not have jurisdiction over
the outer continental shelf.
——October 29: ERDA dedicated its first wind power system at Sandusky, Ohio.
——December 22: Energy Policy and Conservation Act (EPCA) establishes prices
for US crude oil, the Strategic Petroleum Reserve, emergency energy powers for
the president, Corporate Average Fuel Economy (CAFE) standards of 27.5 mpg
by 1985, and cost-effective, appliance energy standards.
——Introduction of catalytic converter mufflers for cars.
1976
——April 5: Congress authorizes production from naval petroleum reserves.
——August 14: Energy Conservation and Production Act (ECPA) creates incentives
for conservation and renewables, funds weatherization for low income homes,
establishes a program for energy standards for new buildings and establishes So-
lar Energy Research Institute (now the National Renewable Energy Laboratory)
to begin the renewable energy era.
——Toxic Substance Control Act reduces environmental and human health risks.
EPA begins phase out of PCBs.
1977
——April 7: President Jimmy Carter indefinitely defers reprocessing of nuclear fuel
and stops construction of Clinch River Breeder Reactor.
——July 5: Solar Energy Research Institute (SERI) opens in Golden, Colorado; re-
named in 1991 as the National Renewable Energy Laboratory (NREL).
——October 1: Department of Energy created from ERDA and the FEA.
1978
——November 9: National Energy Act establishes weatherization grants for low
income families, conservation programs for local governments, energy standards
for consumer products, programs to convert utilities to coal, and energy tax
credits.
——Buried, leaking containers found at Love Canal, NY; clean-up was completed
in 1998.
1979
——March 28: Accident at Three Mile Island nuclear power plant in PA effectively
halts purchase of new reactors in the US.
——Spring: Second oil shock from Iran oil curtailments causes gasoline shortages.
——August 17: President Carter begins to lift price controls on domestic crude oil.
——November 3: US embassy in Iran seized by revolutionaries; Carter suspends
oil imports from Iran on November 14.
P1: GFZ
1980
——April 2: Windfall profit tax on crude oil to give assistance for weatherization
of homes of low income people.
——June 30: Energy Security Act creates the Synthetic Fuels Corporation and funds
renewable energy projects.
——Tax on certain chemicals creates a “superfund” to pay for cleanup when re-
sponsible parties fail.
1981
——January 28: President Ronal Reagan completes price decontrol on domestic
crude oil.
——National Research Council finds acid rain intensifying in northeast US.
——1981–1985: Reagan and Congress debate funding for conservation programs.
1982
——May 24: Reagan proposes transfer of Department of Energy functions to the
Department of Commerce.
1983
——Clean-up of Chesapeake Bay begins.
——EPA encourages homeowners to test for radon.
1984
——December 3: Chemical disaster at Union Carbide plant in Bhopal, India, kills
3000 and partially disables 2700.
1985
——February 4: Department of the Interior reduces estimates for offshore oil (27
to 12 Bbbl) and gas (163 to 91 tcf).
——June 27: EPA modifies mileage test, lowering CAFE’s 27.5 mpg by 2 mpg.
——June 28: EPA curbs tall smokestacks to avoid distant pollution.
——July 16: Appellate Court confirms EPCA’s appliance standards by voiding
DoE’s “no-standard” standard. A key issue was the magnitude of the discount
rate for future benefits.
——Antarctic Ozone Hole for stratospheric ozone discovered.
1986
——April 26: Accident at Soviet reactor in Chernobyl, Ukraine.
——August 21: Large release of carbon dioxide from the depths of Lake Nyos,
Cameron, kill 1800.
1987
——Montreal Protocol for ozone protection bans chlorofluorocarbons.
——December 27: Congress approves Yucca Mountain as the only high-level nu-
clear waste repository under development.
1988
——Congress bans ocean dumping of sewage sludge and industrial waste.
P1: GFZ
1989
——March 24: Exxon Valdez spill of 0.3 million barrels of oil in Prince William
Sound, Alaska.
——March 23: Stanley Pons and Martin Fleishman announce discovery of “cold
fusion.” Physicists do not believe them.
1991
——January-February: Iraq sets 700 oil fires in Kuwait in wake of Gulf War.
——Exxon pays $1 billion for Exxon Valdez spill.
1992
——EPA bans ocean dumping of sewage sludge.
——EPA launches Energy Star Program to identify energy efficient products.
1993
——EPA’s Common Sense Initiative shifts OSHA regulations from pollution-based
to industry-based.
——EPA research finds secondhand indoor cigarette smoke causes 3000 lung-
cancer deaths per year to nonsmokers.
1994
——EPA launches Brownfields Program to restore abandoned city sites.
1995
——70% of US metropolitan areas that had unhealthy air in 1990 meet air quality
standards in 1995.
1997
——EPA restricts particulate matter air emissions down to 2.5-μ diameter.
——Kyoto Protocol to limit greenhouse gases (GHG) is agreed in principle by most
nations, including the US, but the details for trading allotments of GHG are not
certain.
1999
——EPA requires cars, SUVs, minivans, and trucks to have the same emission
standards by 2004, reducing SUV pollution by 77%–95%.
2000
——Most utilities buy combined-cycle natural gas turbines at 55%–60% efficiency,
a major supply side break through, establishing the natural gas era.
2001
——January: California’s partial deregulation contributes to rolling blackouts
caused by many factors.
——July 23: The Kyoto-Bonn Protocol on limiting GHG moved towards imple-
mentation, supported by 178 nations and the EU. Without support from the US,
a former advocate turned critic, the future of the Kyoto process is in doubt. The
US stated it would use voluntary caps on GHG emissions.
——Sales of minivan and SUVs equal sales of cars.
P1: GFZ
2002
——February: President Bush proposed a voluntary 18% carbon reduction by 2012
in terms of greenhouse gas intensity, which is the ratio of national total energy use
of carbon fuels to GDP dollar.
——February 14: Secretary of Energy Spencer Abraham recommends approval of
the Yucca Mountain, NV, geological repository for 77,000 tons of nuclear spent
fuel, to begin in 2010.
——November: Bush Administration relaxes pollution standards on existing coal-
fired power plants and gives managers of national forests more discretion to
approve logging and commercial activities.
2003
——April 1: CAFE standards for light-truck/SUVs rise from 20.7 mpg to 21.0 in
2205, 21.6 mpg in 2006, and 22.2 mpg in 2007.
——June 7: The 18-GWe Three Gorges Dam closes, to be completed by 2009, dis-
placing 1.1 million people.
——August 11: Auto manufacturers no longer contest California’s 2005–20 phase-
in of low- and zero-emission vehicles.
——August 11: Gov. Michael Leavitt nominated to replace Gov. Christine Todd
Whitman as EPA administrator.
——August 14: Electrical grid failure of 62 GWe darkens eight states and Canada
(previously on 11-9-65 and 7-13-77).
——August 27: President Bush gives emission exemptions from the Clean Air Act
for rehabilitated power plants.
2004
——Natural gas prices double and supplies from Canada become more uncertain.
Coal plant orders rise from 2 to 100 and utilities request site approval for 3 nuclear
plants.
——July 9: US Court of Appeals in DC rules against the 10,000 year limit on ra-
diation safety at Yucca Mountain. The court concluded that EPA must either
issue a revised standard that is “consistent with” the NAS peak-dose standard
“or return to Congress and seek legislative authority to deviate from the NAS
report.”
——September 24: The California Air Resources Board required automobiles to
lower carbon emissions by 30%, to be phased in over 2009 and 2016. The Board
claimed the extra cost would be $1000 per vehicle, but it would save $2500 in
fuel. New York and other states claim they will follow suit.
——October 26: General Motors downsizes the Hummer from 6400 pounds (12
mpg) to 4700 pounds (16 city/20 highway).
2005
——January 19: A billion dollar LNG explosion at Skikda, Algeria, clouds plans
for 35 LNG projects in the US.
P1: GFZ
——February 16: The Kyoto Protocol enters into force after Russian ratification.
The 120 nations that ratified emitted 61% of greenhouse gases by signatories,
over the 55% threshold for ratification. The US, which emits 37%, did not ratify,
while India and China did not sign. The carbon-trading allotments initially were
selling for $10/ton.
2006
——October 17: US population passes 300 M. It was 100 M in 1915, 200 M in 1967,
and is predicted to be 400 M in 2043.
P1: GFZ
C. Units
Dimensional Prefixes
10 deka (da) 10−1 deci (d)
102 hecto (h) 10−2 centi (c)
103 kilo (k) 10−3 milli (m)
106 mega (M) 10−6 micro (μ)
109 giga (G) 10−9 nano (n)
1012 tera (T) 10−12 pico (p)
1015 peta (P) 10−15 femto (f)
1018 exa (E) 10−18 atto (a)
Physics Constants
- c = 1/137.036 (fine structure constant)
α = e 2 /h
c = 2.998 × 108 m/s (speed of light in a vacuum)
e = 1.60 × 10−19 coulomb, C (electron/proton charge)
1/4πεo = 9.0 × 10−9 N m2 /C2
εo = 8.8 × 10−12 C2 /N m2 (permittivity of space)
μo = 4π × 10−7 N/A2 (permeability of space)
G = 6.67 × 10−11 N m2 /kg2 (gravitational)
g = 9.807 m/s2 (acceleration of gravity at 45◦ latitude at sea level)
h = 6.63 × 10−34 J s = 4.14 × 10−15 eV s (Planck)
-h = h/2π = 1.06 × 10−34 J sec = 6.59 × 10−14 eV sec
kB = 1.38 × 10−23 J/K = 8.63 × 10−5 eV/K (Boltzman’s constant)
kB T = 0.26 eV = 1/40 eV (at room temperature 300 K)
me = 9.110 × 10−31 kg (electron mass)
mp = 1.673 × 10−27 kg = 1.6726 amu (proton mass)
mn = 1.675 × 10−27 kg = 1.6749 amu (neutron mass)
me c 2 = 511 keV, mp c 2 = 938.3 MeV, mn c 2 = 939.6 MeV
NA = 6.023 × 1023 molecules/gram mole (Avagadro’s number)
volume of mole of gas at STP = 22.4 l at 2.7 × 1019 molecules/cm3
454
P1: GFZ
C. Units 455
Rgas = 8.31 J/g mole K (ideal gas)

σ = 5.669 × 10−8 J/m2 K4 sec (Stefan–Boltzman)
Miscellaneous Useful Numbers

a o = -h2 /mc2 = 0.053 nm (Bohr radius)
e 2 /4πεo = 1.44 MeV fermi (1 fm = 10−15 m)
e 2 /8πεo a o = 13.6 eV (hydrogen binding energy)
E = hc/λ = 1.24/λ (photon energy in eV with wavelength in μ)
Wein λ = 2.90 × 106 /T (blackbody maximum in nm with T in Kelvin)
E = hν = 2 eV (visible light from 6000 K at 0.6 μ)
E = hν = 0.1 eV (IR at 300 K, 12 μ)
band gap: Si (1.1 eV), Ge (0.7 eV)
1.3 fm A1/3 (nuclear radius, A atomic number)
c V ≈ 3Rgas = 24.9 J/g mole K (high temperature specific heat with T θD )
RSI = REnglish /5.67 and USI = 5.67 × UEnglish (insulation)
0 K = −273.15◦ C = −459.7◦ F (absolute zero).
Length
1 mile (mi) = 5280 ft = 1.609 km
1 nautical mile = 1.86 km = 1.16 mi
1 m = 3.281 ft = 39.37 in
1 micron (μ) = 10−6 m = 103 nm = 104 angstrom
1 inch (in) = 2.54 cm
1 fermi (fm) = 10−15 m
universe expanse 1026 m human height 1.8 m nuclear radius 10−15 m.
Area
1 m2 = 10.8 ft2
1 km2 = 0.386 mi2
1 acre = 43,560 ft2 = 1 mi2 /640
1 hectare (ha) = 104 m2 = 2.47 acres
1 barn (b) = 10−24 cm2 .
Volume
1 m3 = 1000 l = 264 US gallons (gal) = 35.3 ft3
1 mile3 = 4.17 km3
1 acre-foot = 43,560 ft3 = 326,000 gal = 1234 m3 = 0.1234 hectare m
P1: GFZ
456 C. Units
1 liter = 1000 cm3 = 0.264 gal

1 bbl petroleum = 42 gal = 0.159 m3 .
Time
1 year = 365.25 days = 8766 h = 3.154 × 107 s ≈ π × 107 s
1 day = 86,400 s
1 shake = 10−8 s
1 age of universe = 4 × 1017 s human 2 × 109 s nuclear 10−23 s (2r /c).
Mass
1 kg = 2.205 pounds (lb) = 32.3 ounces (oz)
1 lb = 16 oz = 453.6 g, 1 oz = 28.4 g
1 metric tonne = 1000 kg = 1.102 tons
1 English ton = 2000 lb = 907.0 kg = 0.907 tonne
Force
1 newton (N) = 1 kg m/s2 = 105 dynes (dyn) = 0.22 lb
Pressure
1 bar (atm) = 76 cm Hg = 760 torr = 14.7 lb/in2 = 105 pascal (1 N/m2 )
Energy/Heat
1 J = 1 N m = 1 W sec = 1 calorie/4.2 = 1 kcal/4200 = 6.242 × 1018 eV = 1 Btu/1055
= 0.738 ft lb
1 eV = 1.602 × 10−19 eV
1 kWh = 3.6 × 106 J = 3412 Btu (electricity at η = 33% uses 104 Btu/kWh)
1 bbl crude petroleum = 5.8 MBtu = (42 gal)(138,000 Btu/gal)
1000 ft3 (STP) natural gas = 100 terms = 1.03 MBtu = 1.09 GJ
1 trillion cubic feet (TCF) natural gas = 1.03 quads, 1 Gt coal = 27.8 quads
1 m3 (STP) = 39 MJ, 1 TCF = 1012 ft3
1 ton coal = 25.2 MBtu = 0.9 tonne coal
1 quad = 1015 Btu = 172 Mbbl = 0.97 TCF = 36 Mt coal = 292 G-kWt h = 1.05 EJ
(1 exaJ = 1018 J)
P1: GFZ
C. Units 457
1 Gbbl = 5.8 quads

1 cubic foot natural gas at STP = 1000 Btu = 1 MJ
1 terawatt-year (TWyr) = 8.76 × 1012 kWh = 31.5 EJ
1 kWh/m2 = 313 Btu/ft2
1 kiloton TNT (kton) = 4.2 × 1012 J = 1012 calories
1 kg fission = 17 kton TNT (60 g/kton)
1 kg DT fusion = 85 kt (12 g/kton)
1 MWthermal day = 1 gram 235 U = 0.3 MWelectric day.
Power
1 W = 1 J/s = 1 N M/s = 1 kg m2 /s3
1 hp = 550 ft lb/s = 0.746 kW = 746 J/s
1 kW = 3412 Btu/h
1 Mbbl/day = 0.365 Gbbl/year = 2.12 quads/year = 71 GWt (thermal)
US 100 quads/year = 3400 GWt = 47 Mbbl/day = 17 Gbbl/year = 100 tcf/year
(equivalent)
US (280 M)/capita = 12 kWt = 60 bbl/year = 12.8 tonne/year coal = 0.35 million
ft3 /year
1 lumen = 1/673 watt visible light
1 lux = 1 lumen/m2 , 1 foot candle = 1 lumen/ft2 = 0.0929 lux.
Air
molecular weight (28.96)
density (1.293 kg/m3 )
sound speed (331.4 m/s)
volume fraction (N2 /78%, O2 /21%, A/0.9%, H2 O/0.4%, CO2 /370 ppm)
specific heat (constant pressure, 1004 J/kg K) and (constant volume, 720 J/kg K)
viscosity (0.17 millipoise).
Water
density [1000 kg/m3 (4◦ C), 997 (25◦ C), 958 (100◦ C), 1025 (salt), 900 (ice)]
latent heat of fusion (333 kJ/kg), vaporization (2.26 MJ/kg)
specific heat water (4.2 kJ/kg K), steam (100◦ C, 2.01 kJ/kg K), ice (2.1 kJ/kg K)
viscosity (17.5 millipoise at 0◦ C, 2.8 at 100◦ C)
flow in Sverdup = 1 M m3 /s
oceans (1350 × 1015 m3 ), ice (29 × 1015 m3 ), ground water (8.3 × 1015 m3 ), lakes
(0.13 × 1015 m3 ).
P1: GFZ
458 C. Units
Earth
radius RE = 6357 km polar and 6378 km equatorial
area = 5.10 × 1014 km2 (oceans 71%)
mass = 5.98 × 1024 kg
g = 9.8 m/s2 (RE /r )2
atmospheric pressure = 105 Pa e−h/H , atmospheric height H = 8.1 km
atmospheric mass = 5.14 × 1018 kg with 1.3 × 1016 kg H2 O, oceanic mass = 1.4 ×
1021 kg.
Sun
solar flux so = 1.367 kW/m2 = 0.13 kW/ft2 = 2.0 cal/min cm2 = 435 Btu/ft2 h
24-h average horizontal flux (40◦ N latitude) = 185 W/m2
mass = 2 × 1030 kg
radius = 0.696 Mkm
distance to Earth = 150 Mkm.
Radiation (colloquial and SI units)

Rate of Decay
1 curie (radiation of 1 g radium) = 1 Ci = 3.7 × 1010 decay/s
1 bequerel (SI) = 1 Bq = 1 decay/s.
Absorbed in Air
1 Roentgen = 1 R = 87 ergs/g = 0.0087 J/kg.
Physical Dose Absorbed
1 rad = 100 erg/g = 0.01 J/kg
1 gray (SI) = 1 Gy = 1 J/kg = 100 rad.
Biological Dose Equivalent (absorbed dose times biological effectiveness Q)
x, γ and e (Q = 1), n (Q = 5–20), alphas and fission fragments (Q = 20)
1 sievert (SI) = 1 Sv = 1 J/kg = 100 rem
1 rem = 0.01 J/kg = 0.1 Sv.
US Annual Average Background Dose (1990 BEIR-V) = 360 mrem (3.6 mSv):
radon (200 mrem), body radioactivity (39 mrem), medical X-rays (53 mrem)
cosmic radiation (31 mrem), sea level (28 mrem), Denver (81 mrem).
Humor Units
1000 aches = 1 MHz, laryngitis unit = 1 Hp, half-bath = 1 demijohn, igloo
circumference/diameter = 1 Eskimo Pi, 1 millionth mouthwash = 1 microscope,
50% large intestine = 1 semicolon, 365 days of low-calorie beer = 1 lite-year,
P1: GFZ
C. Units 459
time between slip on peel to pavement = bananosecond, 8 nickels = 2 paradigms,

build absolutely nothing near anybody = 1 banana, 5 dialogs = 1 decalog, 454
graham crackers = 1 pound cake, 10−6 fish = microfiche, 1 T pins = 1 terrapin,
2 wharves = 1 paradox, 1st step of 1-mile trip = 1 Milwaukee, 10 cards = 1
decards, 16 ft in Twilight Zone = 1 Rod Sterling, tortoise at sound speed =
1 mach turtle, 1 kg falling figs = 10 fig newton, 100 rations = 1 C-ration, 10
millipeds = 1 centipede, 2000 mockingbirds = 2 kilomockingbirds, 1 M bicycles
= 2 megacycles, 365 days = 1 unicycle, 1000 g wet socks = 1 literhosen, 1 M
microphones = 1 megaphone, 3 1/3 tridents = 1 decadent.
P1: GFZ
D. Websites
General
American Geophysical Union: www.agu.org/sci soc/
American Institute of Physics: www.aip.org/history/
American Physics Society: www.aps.org/public affairs
APS Forum on Physics and Society: www.aps.org/units/fps
Congressional Legislation: //thomas.loc.gov
Congressional Budget Office: www.cbo.gov
Congressional Research Service: www.cnie.org/NLE/CRS
DOE Information Bridge: www.osti.gov/bridge/
DOE Labs: www.XX.gov; XX = anl, bnl, lanl, llnl, ornl, pnl, sandia, y12
Economic Report of the President: www.gpoaccess.gov/eop
General Accounting Office: www.gao.gov
Government Printing Office: www.gpoaccess.gov/su docs/
National Academy Press: www.nap.edu
Office of Technology Assessment Legacy: www.wws.princeton.edu/∼ota/
Science: www.sciencemag.org
White House: www.whitehouse.gov
Nuclear Arms
Acronym: www.acronym.org.uk
Argentina/Brazil: www.ABACC.org
Arkin’s Internet: www.sais-jhu.edu/cse
Arms Control Association: www.armscontrol.org
Ballistic Missile Defense Organization: www.acq.osd.mil/bmdo/
Bulletin of Atomic Scientists: www.bullatsci.org/issues/nukenotes/nukenote.html
Carnegie Endowment for International Peace: //ceip.org/field/projects/npp/
npp home.ASP
Center for Defense Information: www.cdi.org
Center for Strategic Education: www.cais-jhu.edu/cse
460
P1: GFZ
D. Websites 461
Central Intelligence Agency: www.cia.gov/public affairs/speeches/speeches.html

Chemical/Biological Defense: www.cbiac.apgea.army.mil/body.html
Chuck Hanson: www.uscoldwar.com
Comprehensive Test Ban: www.ctbt.rnd.doe.gov/ctbt/
CTBT Organization: www.ctbto.org
DoD Cooperative Threat Reduction: www.dtra.mil/ctr/ctr index.html
DoD Defense Threat Reduction Agency: www.dtra.mil
DoD On-Site Inspection Agency: www.dtra.mil/os/os-index.html
Federation of American Scientists: www.fas.org
Garwin’s Page: www.fas.org/rlg/index.html
Harvard: //bcsia.ksg.harvard.edu
Institute for Global Conflict and Cooperation (Univ. California): www-igc.ucsd.edu
Institute for Science International Security: www.isis-online.org
International Atomic Energy Agency: www.iaea.org/worldatom
International Institute for Strategic Studies: www.iiss.org
Monterey Institute International Studies: www.cns.miis.edu
Natural Resources Defense Council: www.nrdc.org/nuclear/default.asp
Nuclear Explosion Monitoring: www.nemre.nn.doe.gov/nemre/
Nuclear Files: www.nuclearfiles.org
Nuclear Threat Initiative: www.nti.org
Pantex: www.pantex.com
Princeton: www.princeton.edu/∼globsec/publications
Russia: www.armscontrol.ru
Science and Global Security: www.princeton.edu/∼globsec/publications/
SciGloSec.shtml
SearchMil: www.searchmil.com
Stanford: //CISAC.stanford/edu
State Department: www.state.gov/t/XX/, XX = isn, pm, vc, trty
Stimson Center: www.stimson.org
Stockholm International Peace Research Institute: www.sipri.se
Strategic Air Command: www.stratcom.mil
Todd’s Atomic Homepage: www.nuc.berkeley.edu/neutronics/todd.html
UK Atomic Weapons Establishment: www.awe.co.uk
Union of Concerned Scientists: www.UCSUSA.org/arms/projects/npp/
resources/
United Nations: www.un.org/Depts/dpa/docs.cdahome.htm
Environment
Acid Rain: //bqs.usgs.gov/acidrain
Air Quality Management District: www.aqmd.gov
British Medical Journal: www.bmj.com
Chernobyl: www.ic-chernobyl.kiev.ua
Congressional Research Service: www.cnie.org/NLE/CRS
P1: GFZ
462 D. Websites
Dr. Everett Koop: www.drkoop.com

DOE Nuclear Waste: //cid.em.doe.gov
DOE Biology/Environment: www.sc.doe.gov/feature/biology and environment.
htm
DOE Carbon Dioxide Information Analysis Center: //cdiac.esd.ornl.gov
Greening Earth Society: www.greeningearthsociety.org
Indiana Law: www.law.indiana.edu/v-lib/index.html
Intergovernmental Panel on Climate Change: www.ipcc.ch
Earth Data: //personal.cmich.edu/∼Franc1m/homepage.htm
Environmental Protection Agency: www.epa.gov
Global Change Research Program: www.usgcrp.gov
NASA: //earthobservatory.nasa.gov
NASA Climate/Radiation: //climate.gsfc.nasa.gov
National Cancer Institute: www.nci.nih.gov
National Center for Atmospheric Research: www.ucar.edu
NOAA Climate: www.ncdc.noaa.gov
NSF: www.geo.nsf.gov/start.htm
Nuclear Waste Technology Review Board: www.nwtrb.gov
Michigan Radiation/Health: www.umich.edu/∼radinfo
Pacific Institute: www.pacinst.org
US Geological Survey: www.usgs.gov
World Bank: //publications.worldbank.org
World Resources Institute: www.earthtrends.wri.org
Energy
American Council for an Energy Efficient Economy: www.aceee.org
American Gas Association: www.aga.org
American Nuclear Society: www.ans.org
American Petroleum Institute: www.api.org
Ballard Fuel Cells: www.ballard.com
Bureau of Transportation Statistics: www.bts.gov/publications/
Clean Energies Future: www.ornl.gov/ORNL/Energy Eff/CEF.htm
Davis Energy Group:www.davisenergy.com
DOE Efficiency/Renewable Energy (600 links): www.eren.doe.gov.
DOE Alternate Energy Vehicles: www.fleets.doe.gov
Efficient Windows: www.efficientwindows.org
Energy Information Agency: www.eia.doe.gov
Energy Star: www.energystar.gov
EPA fuel economy: www.fueleconomy.gov
First Solar: www.firstsolar.com
Fuel Cells: www.fuelcells.org
Hubbert: www.hubbertpeak.com
Hydrogen: www.hydrogenus.org or www.clean-air.org
P1: GFZ
D. Websites 463
Hydrogen Research: www.sc.doe.gov/bes/hydrogen.pdf

International Solar Energy Society: www.ises.org
Lawrence Berkeley National Lab: //enduse.lbl.gov
LBL Buildings: //eetd.lbl.gov/buildings.html
National Renewable Energy Laboratory: www.nrel.gov
National Transportation Statistics: www.bts.gov
Nuclear Power: //web.mit.edu/nuclearpower/
International Energy Agency: www.iea.org
Pacific Gas and Electric: www.pge.com/003 save energy
Princeton: www.princeton.edu/∼cees
Princeton Plasma Physics Lab simulations: //ippex.ppnl.gov
Rocky Mountain Institute: www.rmi.org
US Green Buildings Council: www.usgbc.org
Windpower: www.windpower.org
P1: GFZ
E. Symbols
Variables and variable parameters in Roman letters are displayed in italics. Param-
eters represented by Greek letter and the physics constants are not displayed in
italics. Variables that represent volume or surface densities are denoted in lower
case, while final values and quantities are often displayed in UPPER CASE. How-
ever, there are exceptions pursuant to the normal traditions of the discipline. Vectors
are displayed in bold type only when useful. Note that separate entities may be rep-
resented by the same symbol. The distinct meanings are apparent in the chapters
where they appear.
E.1 Chapters 1–5: National Security

A = atomic number fraction of sites in violation, N/S;
a = acceleration isotopic abundance, 0–1 for 235 U/U
B = laser brightness, w/steradian; f , i, o = focal length, image/object dis-
magnetic field, Tesla (T) tances
C = a constant; fl = fluence, total particles or energy per
cost of an item, $ area
Cd = drag coefficient GWe = 109 W of electrical power
CEP = circular error probable radius GWt = 109 W of thermal power
(50% hits), m or nmi g = acceleration of gravity, 9.8 m/s2 , or
c = specific heat, J/kg-K 32.2 ft/s2
D = diffusion constant; mirror diame- g, h c , s = image, composite point-spread
ter; function, and signal
infrared detectivity; thickness G, H, S = Fourier transforms of the
E = energy, J, eV, or Btu above
F = force, newtons, or pounds H = hardness of silos, psi, or atmo-
F, P, W = feed, product, waste (tails) spheres;
enrichment flows thickness of plate
f = frequency (ω = 2πf); Isp = specific impulse in seconds
flux of energy, W/cm2 ; (Vexhaust /g).
464
P1: GFZ
E. Symbols 465
i e = irradiance, power density or flux, R = reliability, 0–1;

W/cm2 missile range, km
k = neutron multiplication factor in a RU = ratio of isotopes, often 235 U/239 U
particular geometry rc = critical distance for a weapon effect
kB = Boltzman’s constant S = number of sites being inspected;
kWe = 103 W of electrical power entropy, J/kg-K, or J/mole-K
kWt = 103 W of thermal power SSKP = single-shot kill probability with
L = lethality, a function of CEP accuracy 100% reliability
and Y yield SWU = enrichment separative work
L m , L v = latent heats of melting, vapor- unit, J/kg
ization, J/kg t = time
m = mass of a particle or sample T1/2 = half-life = 0.693τ (τ is mean life)
mb = seismic magnitude for body pres- T = temperature, K, ◦ C, ◦ F
sure wave V = violations to a treaty; voltage; vol-
n = density of neutrons or fissile mate- ume;
rial; value of feed, waste, or product
number of inspections per year; vshock = shock velocity
number of microstates for enrich- Y = nuclear yield, kton or Mton
ment; Ymod = Young’s modulus
Nn = number of neutrons, gamma rays, ----------
etc. α = optical absorption (1/e) length
Nx = strength of conventional forces for β = Earth latitude;
nation x ballistic missile coefficient, Pascals
Pkill-n = probability of kill with n- ε = effectiveness of conventional war
attacking warheads fighting
P = power, watts (W) η = efficiency of process
p = pressure, Pascals, psi, or atmo- number of neutrons emitted/
spheres; absorbed neutron by all
radial probability of missile de- γ = cp /cv for a gas; 1/(1 −v2 /c2 )1/2
struction λ = wavelength
Q = heat energy λpath = mean free path
q = electric charge, coulomb ν = neutrons emitted/neutron capture
Rc = critical mass radius by fissile isotope
RD = radius of deuterium or tritium θ = missile angle above horizontal; φ =
Rgas = ideal gas constant azimuthal
Riso = ratio of two isotopic concentra- ρ = density of mass, neutrons, etc.
tions σ = nuclear cross section
Rrefl = reflectivity, 0–1 τ = mean life; exponential rise time
E.2 Chapters 6–9: Environment

A = area of pollution basin Bo = magnetic field amplitude, T or
a = albedo, fraction of sunlight reflected Gauss
from a planet b = breathing rate, m3 /sec
P1: GFZ
466 E. Symbols
C = heat capacity for a system, J/K; N = total number of moles or atoms

carbon production, tonnes/year NACH = number of air-exchanges per
c = pollution concentration, kg/m3 ; hour in a building
specific heat density, J/kg-K NC = net carbon production rate from
c p , c v = molar specific heat, J/mole-K natural causes
(pressure, volume) Nozone = number of ozone molecules in
D = diffusion constant, m2 /s the stratosphere
Drad = radiation dose, Sv or Rem n = scaling parameter for cities, n =
dd = degree days in a time period, ◦ C- L/L o ;
day or ◦ F-day number of atmospheric IR absorp-
E = energy, J, Btu; tion layers
expected value of disease for P = amount of pollution, kg, ton,
nonexistent cause Curies;
E, P = evaporation, precipitation, population in a plume or on Earth
tons/yr P1 = failure probability for task 1
E rms = root-mean square electric field, p = osmotic pressure increment
volts/m p = % chance that additional risk is ran-
E cap = energy use per person dom
e d = demand elasticity, d/d divided pH = −log[H ion moles/liter], acidity
by p/ p (price) measure
F = flow rate, kg/s pk = reaction dissociation constant in
f = pollution flux, kg/m2 s; 10− pk
energy flux, W/m2 R = electrical resistivity, ohms;
f forcing = atmospheric radiative forcing reflection coefficient, 0–1
flux, W/m2 Rozone = ozone destruction rate
f IR = fraction of Earth’s surface emis- RR = relative risk from exposure vs.
sions in the IR nonexposure
grad = ground radiation rate, Ci/m2 Q = heat energy, J, Btu;
H = inversion height from rising atmo- relative biological effectiveness, 1–
spheric temperature. 20
Iind = pollution inhaled by an individ- S = pollution flow from production or
ual transfer, kg/sec
Ipop = pollution inhaled by all persons so = solar flux above Earth’s atmo-
in a plume sphere, 1.37 kW/m2
Isolar = solar ozone production rate Ta , Ts = temperature at top of atmo-
K H = aqueous/atmospheric concentra- sphere and on surface
tion, moles/liter-atmos. T = transmission coefficient, 0–1
ktherm = thermal conductivity u = wind velocity, m/sec
L = length of side of a city or pollution V = volume of atmospheric box (area ×
plume. height)
LL (UL) = lower (upper) limit for 95% Vlife = value of a life
likelihood vdep = deposition velocity from Stokes’
L evap = latent heat of vaporization law
Mwt = molecular weight, summed [x] = concentration of ion x, moles/liter
atomic weight, grams ----------
P1: GFZ
E. Symbols 467
α, β = coefficients linking radiation and λforce = ratio of radiation forcing to sur-

cancer face temperature rise
ε = emissivity of surface or air layer, μ = chemical potential increment
0–1 ρ = matter density
= lapse rate temperature reduction, σ = population surface density;
6.5 K/km standard deviation; plume half-
η = furnace efficiency; energy efficiency; width;
viscosity electrical conductivity = 1/ρ =
ηC = useable energy per unit carbon 1/ohm-m
consumption θ = angle of sun from zenith;
λ = growth or decay rate plume angle width
E.3 Chapters 10–16: Energy

B = benefit from an investment, $ F = fuel economy in miles/gallon in
b,d = birth and death rates CAFE regulations
b = breathing rate Faero , Froll = drag and rolling forces
C = cost of an investment f = flux of particles or energy,
c = specific heat, J/kg-K number/m2 s or W/m2
c p , c v = molar specific heat, J/mole-K f C = fraction of Carnot efficiency
(pressure, volume) f int = internal heat flux in buildings,
C = heat capacity, J/K W/m2
C, c = cost of an item, or account balance G = gasoline consumption rate,
Cd , Cr = aerodynamic and rolling drag Mbbl/day
coefficients GWe = 109 W of electrical power
CCE = cost of conserved energy from an GWt = 109 W of thermal power
investment h = elevation above sea level;
CRR = capital recovery rate (interest enthalpy = internal energy + pV, J
plus capital, also α) or Btu
d = discount rate hp = horsepower
d, s = quantity of demand and supply I = moment of inertia;
of a product integrated solar flux, kwh/m2 -day,
dB = decibel sound level Btu/ft2 -day;
dd = degree days in a time period, ◦ C- sound intenisty, W/m2
day or ◦ F-day i = interest rate
E = energy, J or Btu kWe = 103 W of electrical power
E gap = semiconductor band gap, eV kWt = 103 W of thermal power
e d = demand elasticity, d/d divided L = number of laborers
by p/ p(price) LCC = life cycle cost with all the costs
e s = supply elasticity, s/s divided by M = maintenance rate
p/ p M, m = mass
e 12 = cross elasticity, d/dof fuel 1 di- NPV = net present value (initial cost +
vided by p/ p of fuel 2 PVB of benefits)
P1: GFZ
468 E. Symbols
N = number of conservation measures so = solar flux above atmosphere, W/m

per year or Btu/ft2 -hr
n = volume density of matter or parti- T = transmission coefficient;
cles payback period;
P = population; tensile strength
power, watts (W) or horsepower Thot ,Tcold = hot, cold temperature for
(hp); Qhot ,Qcold
fixed payment total rate (CRR × in- Tfree = free temperature from thermal
vestment cost) power in a building
Prod = productivity = GDP/labor-force T1/2 = half-life
PVC , PVB , PVnet = present value of cost, T2 = doubling time for exponential
benefit or net benefit growth, 70/λ(%/time)
p = price of product; U = thermal transmittance =
pressure, Pa, psi; 1/R,m2 -◦ C/W, ft2 -hr-◦ F/Btu
momentum, kg m/s vsound = speed of sound
p = stack pressure from wind and ther- Wpeak , Wt , We = peak, thermal and elec-
mal differences tric power, Watts
Q = heat energy, J or Btu ----------
Qhot , Qcold = heat transfer from hot and α = lossiness, heat loss rate = αT;
cold temperatures solar collector absorbtivity
Q∞ = total recoverable resource at a φ = inflation rate beyond normal infla-
fixed cost/unit tion
Qo = initial quantity of resource con- η = energy efficiency, Wnet /Qin
sumed = time half width of modern
Qm = quantity consumed at maximum petroleum era
pumping rate η = economic efficiency, GDP/energy
q = quantity of a product used
R = thermal resistance (1/U), W/m2 -◦ C, λ = growth rate
Btu/ft2 -hr-◦ F σDT = DT-fission cross-section
RACH = air exchange rate in air σ v = average cross-section × velocity
changes/hr (ach) at 5 keV
RB = benefit rate, $/year μs , μk = static and kinetic coefficients of
RDT = DT-fusion reaction rate friction
RE = radius of Earth; θ = angle of sun from the zenith
energy rate, Mbbl/day ρ = density of air or material
RSI = RUS /5.67 = 1/USI = 1/5.67UUS τ = thermal decay or rise time;
r = compression ratio, Vinitial /Vfinal torque with angular variables θ,
s = supply of product; ω, α
P1: GFZ
F. Glossary
Chapters 1–5: National Security

ablative shock—x-rays penetrate a short distance, vaporize a small surface volume,
giving a shock to the object.
ABM—the Anti-Ballistic Missile Treaty constrained US/Soviet strategic defenses
from 1972 until 2002.
adaptive optics—submicron mirror movements can improve some monitoring and
defense technologies.
adaptive preferential targeting—SDI targeting that maximizes defense effective-
ness if all incoming trajectories are known.
anti-satellite weapon—easier to attack satellites than missiles, but threatens all
nations that have satellites.
anti-submarine warfare—ASW consists of ocean sensors, trailing submarines,
depth charges, and homing torpedoes.
bare sphere—the critical mass of a spherical fissile weapon without reflectors,
tampers and compression.
Becker nozzle—separates isotopes by passing a gas through a tiny tube around a
sharp corner.
bhangmeter—an optical detector in orbit to detect nuclear explosions and deter-
mine their yields.
bomblet—10 kg reentry vehicles, released at the end of boost phase, can carry
biological weapon materials.
boost-phase—first minutes of powered flight of a missile’s trajectory before mid-
course/MIRV and reentry.
boosted primary—contains 2 H and 3 H to increase efficiency of fission with a neu-
tron pulse that skips neutron generations.
breakout—massive violation of a strategic arms treaty that could destroy the un-
suspecting treaty partner.
469
P1: GFZ
470 F. Glossary
breeder reactor—a more compact reactor that breeds more Pu fission fuel than it
consumes.
burn a hole in the atmosphere—the heating of a tube of atmosphere to reduce air
density, to allow an electron beam to pass.
chaff—a countermeasure to spoof the defense, small metallic wires that reflect
radar.
countermeasures—protection of attacking missiles by blocking the defense’s sen-
sors.
coupled explosion—a nuclear explosion is joined to the earth by tamped soil/rock
(see decoupled explosion).
CTBT—Comprehensive Test Ban Treaty bans all nuclear explosions for-
ever/anywhere; defeated by US Senate in 1999.
critical mass—amount of fissile material that produces slightly more neutrons than
are lost by absorption or leakage.
cookie-cutter—the simplified zero/one function for silo destruction.
decoupled explosion—a small nuclear explosion in a large cavity; observed yield
can be reduced up to a factor of 70.
denatured uranium—adding natural uranium to weapons grade 235 U to remove
its weapon capability; cannot be done with Pu.
diminishing returns—reduced destruction per warhead with increased numbers
of warheads.
DoD and DOE—US Departments of Defense and Energy.
downblend U—mixing of 90% enriched U with natural or 1.5% enriched to obtain
4.4% reactor fuel.
download warheads—reducing the number of MIRVed warheads, such as Min-
uteman’s reduction from 3 to 1.
DPS—Defense Program Satellites at GEO with IR sensors to detect launches of
missiles.
drawdown curve—graph of the number of surviving warheads as a function of
the number of attacking warheads.
duty factor—defensive weapons in orbit have a duty factor of 5–10% above the
opposition’s warheads.
earth-penetrating warhead—B61/11 bomb; increases coupling to earth, but a new
5-kton version was rejected by the congress.
effective verification—ability to respond to cheating before significant military
loss, and to observe smaller cheating.
endoatmospheric—a location above the atmosphere, at an altitude above 10 km,
approximately.
EMP—electromagnetic pulse caused by asymmetrical currents or by electrons or-
biting in Earth’s magnetic field.
P1: GFZ
F. Glossary 471
exoatmospheric—a location in the atmosphere.

first strike—a pre-emptive start to nuclear war
fission—the splitting of 235 U or 239 Pu into two fission fragments, yielding
17 kton/kg.
fluence—total number of particles or energy traversing a unit area.
flux—flow of particles, fluid or energy per unit area/second
fratricide—the destruction of one warhead by another warhead from the same
side.
fuel cycles—once-through fuel cycle uses nuclear fuel without reprocessing; closed
cycle entails reprocessing to obtain Pu.
fusion—the combining of 3 H and 2 H to give He and a neutron, yielding
85 kton/kg.
GEO—geosynchronous orbit at 40,000 km radius, stationary with respect to Earth
for DPS and other satellites.
GPS—Global Positioning Satellites allows missile targeting to within a meter, at
26,600 km radius.
gravitational bias—the 0.3% oblateness of Earth must be considered for high-
accuracy missile trajectories.
hard target—silos or command centers with a hardness of 1000 to 5000 psi.
Hiroshima—city in Japan, targeted by the Thin Man, gun-type, uranium bomb,
with 13-kton yield, killing some 75,000.
HEU—highly enriched uranium with over 20% 235 U (usually 90%), which is not
weapons usable below some 50% enriched.
hydrogen bomb—a bomb that uses radiative compression from the fission primary
to heat 2 H–3 H to fuse in the secondary.
IHE—insenstiive high explosives that make warheads safer, but which take up
extra space, compared to HE used on SLBMs.
implosion—quickly assembles a critical mass by explosion to avoid preinitiation
from spontaneous neutrons.
irradiance—power per unit solid angle in watts/steradian.
kill probability—silo destruction depends on hardness, yield, accuracy, reliability,
warhead numbers, bias, etc.
KKV—kinetic kill vehicles collide and destroy missiles and reentry vehicles in
midcourse phase.
Lanchester, Frederich—developed conventional-arms battle models, favoring
large forces by strength squared.
launch on warning—an option in the SIOP to launch missiles, when electronic
sensors perceive we are under attack.
launch weight—typically 20–30 times the throw-weight of a missile.
P1: GFZ
472 F. Glossary
layered defense—multiple layers in boost, midcourse and reentry phases to en-

hance kill probability.
LEO—low Earth orbit at altitudes of 100 to 1000 miles with a revolution period of
1.5 hours.
LEU—low enriched uranium with 235 U less than 20%.
lithium deuteride—a stable solid that contains deuterium and produces tritium in
a neutron bath for quick-assembly fusion.
MAD—Robert MacNamara’s mutual assured destruction would destroy 25% of
population and 50% of industry of US or USSR.
minimum deterrence—Robert Macnamara’s assured destruction of Soviet popu-
lation (25%) and industry (50%).
Minuteman vulnerability—worst case scenario was 80–90% vulnerable, but only
part of the triad.
MIRV—multiple, independently targetable reentry vehicles with 10 RV’s per SS-
18/Peackeeper.
MOX— for light water reactor fuel, combination of 5% reprocessed Pu with 95%
natural uranium.
Nagasaki—city in Japan targeted by the Fat Man plutonium-implosion bomb with
a 22-kton yield.
neutron bomb—1–10 kton bomb with 50% fusion uses fast neutrons to kill tank
drivers quickly but buildings survive.
NWS/NNWS—under NPT, big-five nuclear weapon states (NWS) may have nu-
clear weapons and the 180 non-NWSs may not.
NPT—the 1970 Nuclear Nonproliferation Treaty is the controlling legal document
for weapon possession and safeguards.
nuclear winter—theory that massive burning from a nuclear attack lofts carbon
into the stratosphere, cooling planet Earth.
off-alert—to minimize accidental or mischievous attacks, delaying technologies
are inserted to prevent rapid responses.
Okla, Gabon—location where 1.8 billion years ago, a natural reactor burned
100,000 years, lowering 235 U content to 0.4%
overpressure—2000 psi destroys most silos and 5 psi destroys most buildings.
popup missiles—based on submarines close to the attacker, difficult to achieve in
hostile regions.
primary—the first stage of a nuclear weapon from fission of 239 Pu or 235 U in gun
or implosion weapons.
quick from the dead—Bernard Baruch’s 1946 charge to the UN to control nuclear
weapons before they proliferate.
railgun—based in space and uses electromagnetic forces to attack reentry vehicles.
P1: GFZ
F. Glossary 473
reactor-grade Pu—uranium fuel rods that remain a long time in a reactor having
over 20% 240 Pu.
reentry vehicle—a warhead packaged in heat resistant materials to allow reentry
into the atmosphere.
Richardson, Lewis—developed action–reaction spending model for nations fear-
ing their national security.
rocket equation—Newton’s second law combining exhaust gas thrust with vari-
able mass; can include gravity and drag.
SALT/START/SORT—strategic arms control treaties in force between the US and
USSR/Russia between 1972–2013.
scaling laws—useful to understand the basic physics and relative sizes of nuclear
warheads.
SDI—President Ronald Reagan’s Strategic Defense Initiative, 1983–1990; elements
are still being researched.
secondary—ignited by the fission primary, it contains 6 LiD and uranium to give
both fusion and fission energy.
second strike—a response to a first strike, for which the US exceeds the minimum
deterrence level.
sine qua non—a necessary condition, such as space-based weapons are a sine qua
non for a successful SDI.
SIOP—standard integrated operation plan, the US nuclear war plan.
SNM—special nuclear material encompasses the weapons isotopes of 235 U, 232 U,
and Pu.
soft target—buildings with hardness of 5–100 psi overpressure.
specific impulse Isp —rocket fuel exhaust velocity divided by g, with units of sec-
onds.
spontaneous neutrons—spontaneous fission of 240 Pu gives spontaneous neutrons
that reduce the yield of some weapons.
strategic weapons—nuclear weapons on ICBMs, SLBMs, and heavy bombers with
ranges over 5500 km.
SWU—separated work unit, the amount of enrichment work needed to raise en-
richment levels of uranium.
throw-weight—typically 3–5% of an ICBM/SLBM’s launch weight.
theater ballistic missile—capable of 2000-km range, less than intermediate range
5500 km (SS-20 and Pershing-II in INF).
time line—the critical time for decisions and operations for ABM systems.
trajectory, flat Earth—this simplification retains much of the physics of trajectories.
trajectory, spherical Earth—trajectories in closed form, with minimum energy at
22◦ above horizon for 10,000-km range.
P1: GFZ
474 F. Glossary
trajectory, oblate Earth—this model requires Runge-Kutta numerical solutions, or

perturbation approximations.
triad—three nuclear legs consisting of ICBMs, SLBMs and heavy bombers, plus
theater weapons.
tritium—3 H fuses with 2 H in boosted primaries or hydrogen secondaries, with a
12-year half life.
upload warheads—to increase number of warheads on an ICBM.
weapons-grade HEU/Pu— uranium enriched to over 80% and Pu with less than
6% 240 Pu (but all Pu can be used).
weapons yield—varies from 1 lb to 58 Mton (which was partial yield of a 100–150
Mton Soviet weapon).
x-ray laser—laser that is pumped with a nuclear weapon, pops-up from close-in
subs, but difficult to do successfully.
yield excursion—a beginning nuclear weapon state may not be able to predict the
yield of its explosions.
Chapters 6–9: Environment

acid rain—contains H2 SO4 and HNO3 at pH’s between 2 and 5.
allowance trading—transfers pollution mitigation dollars to a more cost effective
use, in lb sulfur per MBtu energy.
albedo—the portion of solar energy reflected from a planet
anthropogenic—mankind’s impact on nature, as from the increase in CO2 and CH4
which can affect climate.
association—epidemiology uses association (not correlation) when linking an ef-
fect with a possible cause.
atmosphere (top)/surface temperature—Earth (255 K/287 K) versus Venus (229
K/750 K) from CO2 and albedo differences.
atmospheric layers—useful to display the physics and get approximate results,
but not as good as GCMs.
BEIR—Biological Effectiveness of Ionizing Radiation standards of the National
Research Council.
B to E fields—time varying B fields induce E fields in the body, but these are usually
less than those from thermal motion.
BWR—boiling water reactors use steam directly from the pressure vessel to drive
turbines.
Btu/Joule tax—a carbon tax based on carbon used, a fairer approach than a tax
based on cost.
P1: GFZ
F. Glossary 475
carbon tax—could discourage carbon usage, but the market forces are unsure.
case/control—epidemiological analysis uses a 2 × 2 matrix of cases/controls and
exposed/nonexposed.
cell phone—a standard of 1 W/kg near the ear is based on heating of flesh.
CFC—chrlorofluorocarbons are primarily responsible for reduction of the ozone
layer.
closed fuel cycle—spent fuel is reprocessed, to obtain plutonium that is used in
breeders or thermal reactors.
confounding bias—apparent, but absent, connection between ill health and a vari-
able, but it is only a correlation.
Debye temperature—temperature at which higher energy vibration modes are
excited with c v = 3Rgas .
displacement current—this plus real current equals total current.
Dobson unit—the number of ozone molecules above 1 cm2 is the Dobson unit
times the molecules/cm3 at STP. 1 Dobson = 2.7 × 1020 O3 /m2 .
ECCS—emergency core cooling system replaces the lost coolant in a core if a pipe
breaks.
ELF–EMF—extremely low frequency electromagnetic fields that some conjectured
could cause cancer.
EPA—US Environmental Protection Agency.
evaporation/precipitation—transfers 40,000 km3 /year from ocean to land as rain.
faculae—bright spots on the sun that radiate ultraviolet that might effect climate.
fault tree analysis—the probability of accidents with airplanes and reactors are
estimated with parallel and series logic.
feedback—clouds can enhance or diminish temperature changes from the primary
effects of greenhouse gases.
Fick’s law—the flux of impurities is proportional to the gradient of its concentra-
tion, f = −D∇c.
GCM—general circulation model uses physics and empirical data to estimate cli-
mate changes.
GCRP—US Global Change Research Program helps coordinate and evaluate US
climate change research.
grandfathered—older facilities are often given delays or some exemptions for new
environmental regulations.
GHG—greenhouse gases—CO2 , CH4 , CFC, N2 O, O3 —caputre infrared and raise
the surface temperature.
ICRP—International Commission on Radiological Protection.
inversion level—rising warm gases cool until they no longer rise at the inversion
level.
P1: GFZ
476 F. Glossary
heat islands—local temperature rise up to 6◦ C from big city urbanization, requiring

additional air conditioning.
Hill’s criteria—the criteria to judge the degree of association between health effects
and possible causes.
hormesis—the controversial theory that small doses or radiation are beneficial.
HTGR—high temperature gas reactors which use carbon for a moderator, which
has more specific heat but is flammable.
IPCC—Intergovernmental Panel on Climate Change; a global organization.
Kyoto Protocol—a 1997 global agreement to set limits on CO2 production, using
allowance trading mitigation.
lapse rate—rising warm air cools as it expands and evaporates moisture content
at 6.5 K/km.
LOCA—loss-of-coolant accident; the melting of a core from uncooled beta-decay
heat from fission fragments.
LWR—light water reactors use 3–5% enriched uranium and light water
coolant/moderator in PWRs and BWRs.
HWR—heavy water reactors use natural uranium and heavy water
coolant/moderator; more proliferation concern than LWRs.
magnetosomes—speculation exists that single-crystal magnetic particles; oscillat-
ing in B-fields, could cause cancer.
maximum contamination level—the allowed level of water contamination under
EPA’s primary standard.
mean life—the average life of radioactivity, which is 44% longer than its half-life.
Milankovitch—showed that Earth has cycles of 20k, 40k, 100k, and 400k years.
Montreal (1987) and London (1990) Protocols—ban production of chloroflorocar-
bons, CFCs.
NRC—US Nuclear Regulatory Commission.
NAAQS—EPA’s National Ambient Atmosphere Quality Standards for CO, NO2 ,
O3 , SO2 , Pb, and particulates.
18
O/16 O, 2 H/1 H—ratios used to determine past temperatures; heavier isotope con-
tent rises in ice at higher temperatures.
ocean accoustical tomography—temperature-dependent wave velocities give time
spectrum to determine ocean temperature.
once-through fuel cycle—reactor spent fuel is not reprocessed, but stored in a
repository.
particulate matter—particles with diameters more than 10 and 2.5 μ are controlled
under the Clean Air Act.
passively safe—reactors that remain totally safe when loosing power or coolant,
true in principle, but hard to do.
P1: GFZ
F. Glossary 477
person-sievert—units of effective radiation dose received by a population (1 person

Sv = 100 person-Rem).
pH acidity—log[H ion moles/liter]; from 0–7 is acidic, 7 is neutral, 7–14 is basic.
plumes—SO2 , 137 Cs, and Pu plumes depend on wind velocity/turbulence, source
strength, deposition velocity and response.
Primary Air Quality Standard—set by EPA for CO, NO2 , O3 , SO2 , Pb, and partic-
ulates.
potassium iodide—pills of KI will remove radioactive iodine from the body for
those near a reactor accident.
probabilistic risk analysis—PRA uses fault tree analysis with serial and parallel
logic to quantify risk.
p-value—observed significance level; probability that health problem is a random
occurrence rather than due to a cause.
PWR—pressurized-water reactor uses a heat exchanger to transfer heat from the
reactor to make steam to drive turbines.
radiation doses—energy absorbed (1 gray = 100 rad) and effective dose in
(1 sivert = 100 rem).
radiative forcing—adds downward flux of energy from CO2 , H2 0, CH4 , clouds
(+/−), ice reduction, etc., in W/m2 .
radon daughters—218 Po, 214 Po, 214 Pb and 214 Bi from decay of 222 Rn with a 3.8 day
half-life.
rational actors—the assumption that nations and citizens decide issues rationally.
reflected plume—the plume that hits the ground bounces up and is added to the
rest of the plume.
relative biological effectiveness—a factor between 1 and 20 that takes into account
the type of radiation.
relative risk—disease rate of those who have been exposed divided by the rate of
those that have not been exposed.
rem—Roentgen equivalent man; radiation dose of 0.0087 J/kg times a RBE factor
for the type of radiation.
replacement time—the time for the fluid to turn over (mass/flow rate) or blow
away (length/wind velcoity)
right-of-way—limits to constrain construction are often set on the basis of E/B field
strengths from powerlines.
scrubbing—the removal of SOx with crushed limestone CaCO3 and moisture.
sequesting of CO2 —storing CO2 to prevent its entry into the atmosphere; small
leakage is necessary.
Smog—the combination of NOx , hydrocarbons and sunlight creates ozone in urban
air.
P1: GFZ
478 F. Glossary
Tambora—volcano that exploded 80 km into the atmosphere in 1815, cooling Earth

by 1◦ C for a few years.
thermal rise time—either the exponential rise time or the time it takes for a system
to exceed its safety limits.
thermal efficiency—the ratio of useful output to energy input.
ultra violet purification—bare fluorescent bulbs can help save 3.8 million
deaths/year deaths that result from bacteria in water.
WAIS—West Antarctic Ice Shelf rests on firm earth and could be dislodged, raising
ocean levels.
warhead accident—non-nuclear accident; could spread Pu, but precautions have
been taken; linked to nuclear testing issue.
Yucca Mountain—a geological repository north of Las Vegas for US spent fuel from
commercial reactors.
Chapters 10–16: Energy

adiabatic—compression or expansion without heat transfer, because of short du-
ration of the process.
aerogel—tiny, 10-nm fused silica spheres with large convection and conduction
resistance, which can be used with low E.
air-to-air heat exchanger—a heat exchanger that transfers from exiting heat to
incoming air, while removing pollution.
astronomical unit (AU)—distance from Earth to sun of 150 million km (93 million
miles).
avoided cost—the cost utilities must pay (with corrections) to alternate electrical
suppliers under PURPA
balance point—the outside temperature where a building has good room temper-
atures, usually 68◦ F, without a furnace.
blower door—air leaks can be found and plugged by slightly reducing the internal
pressure in a building.
bridging fuel—natural gas is the most likely, clean and cheap fuel, but how much
is there?
building energy standards—California’s laws led the nation, saving over 50% in
energy in new buildings.
busbar cost—the cost electricity at the power plant is about 3 c/kWh cheaper
before transmission and distribution.
CAFE—the 1975 Corporate Average Fuel Economy standards raised cars from 13
to 27.5 miles/gal; what is next?
P1: GFZ
F. Glossary 479
capacity factor—fraction of time that a power plant of any type is operating at

rated capacity.
capital recovery fee—higher fees for those who want more electrical power; has
been done with water hookups.
capital recovery rate—the rate paid (%/year), based on initial investment, interest
rate and time period of the loan.
clock thermostat—allows lower temperature at night by turning the furnace on
before one gets up.
compact fluorescent—cost $1–2 and save 75% of electricity, as compared to incan-
desent bulbs.
coal—a solid at 7000 to 12,000 Btu/pound contains 20–90% carbon and 0.6–4.3%
sulfur.
cohort—demographic analysis divides populations into similar groups; such as 10
cohorts of 10 years each.
constant/current dollars—constant dollars have inflation removed, while current
dollars are the ones we spend.
combined-cycle gas turbine—the big supply breakthrough since the oil embargo,
two cycles giving 55–60% efficiency!
cost of conserved energy—the investment to save energy is prorated over the
energy saved, giving a CCE of c/kwH, $/gal.
cycles—Carnot and its children; Brayton, Diesel, Otto, Rankine, combined, super-
heat and more.
decibel level reduction—road and plane noise is diminished with geometric
spreading and absorption at 7 dB/km.
degree days—prediction of heat loss; proportional to coldness outside (T) times
length of time under consideration (t).
discounted benefit—$1 today is better than $1 10 years from now; future benefits
must be discounted to the present.
duty/capacity factor—the fraction of time that a power plant is operating at full
power.
elasticity—the change in supply or demand from changes in price.
elastic/inelastic—elastic markets respond sharply to price changes, while inelastic
ones are sluggish; both exist.
enthalpy—energy available for thermodynamic processes, h = internal energy +
pV.
feebates—a plan to pass money from those who buy consumptive cars to those
who buy conserving cars; as revenue neutral.
fleet average—CAFE energy regulations require manufacturers to average fuel
economy over their fleet of cars.
P1: GFZ
480 F. Glossary
fuel cell—combines hydrogen and hydroxide to make electricity/water from hy-

drogen or methanol/gasoline with help.
fuel economy—energy use in miles/gallon; the inverse of the gasoline consump-
tion rate.
free temperature—a typical house is 1–2◦ C warmer inside because of body heat
and machines, it can be 10–20◦ C
Freedom Car—President George W. Bush’s hydrogen, fuel-cell car to lower gasoline
consumption.
gasohol—Brazil motor fuel with ethanol from sugar is more easily produced than
US gasohol from starch.
heat pump—heat goes from the colder to the hotter, not by itself, but with an electric
heat pump.
hidden hand—Adam Smith’s marketplace keeps prices low with self interest look-
ing for good buys
high-frequency ballast—saves 10% of plasma discharge losses in fluorescent tubes
when run at 30 kHz.
hybrid car—the Prius led the way with an electric motor and IC engine in parallel,
towards a more optimal, smaller size IC.
indirect cost—increased imports of oil will have a wide variety of indirect economic
effects.
ICF—inertial confinement fusion contains hot plasma fuel only because it fuses
quickly,
indirect solar flux—diffuse scattering from clouds gives 10% or more of the direct
solar flux.
infiltration—warming incoming cold into a building uses 25% of total heating
energy.
integrated solar flux—the daily integrated solar flux is about 2000 Btu/ft2 or
5 kWt h/m2 .
King Hubbert—US geologist who correctly predicted US petroleum production
would peak in 1970 at 9 Mbbl/day.
life cycle costs—the total costs/benefits discounted to the present; good policy
minimizes life-cycle costs.
load- and skin-dominated—large buildings need less insulation; houses need more
insulation and are move affected by climate.
lossiness—energy loss rate of a building is equal to the lossiness factor times the
temperature difference.
low-E window—special materials on glass reflect, rather than absorb, infrared at
wavelengths longer than 1.5 μ, below ωplasma .
Lawson criterion—product of confinement time and particle density/cm3 for en-
ergy break-even (with 1/3 recovery heat).
P1: GFZ
F. Glossary 481
LED—light emitting diode, an efficient, but yet expensive light source.

LNG—liquified natural gas can be transported to places where it could otherwise
not be used.
modal shift—some of US energy savings are due to a changed society, from steel
mills to silicon valley.
National Ignition Facility—200 beam laser with 1.8 MJ pulses for stockpile stew-
ardship at Livermore.
natural gas—at 1000 Btu/ft3 contains 75% methane CH4 , 6% ethane C2 H6 , 4%
propane C3 H8 and 12% uncombustible.
OPEC—Organization of Petroleum Exporting Countries; its fraction of world oil
production is projected to rise to 50% by 2020.
OTEC—Ocean Thermal Energy Conversion, an unlikely source of power from the
ocean’s 20◦ C temperature difference.
payback period—the amount of time for conservation savings to recover the orig-
inal investment.
peak watt—photovoltaic output when solar flux is a maximum of 1 kW/m2 , when
sun is in the zenith on a clear day.
petroleum—a liquid at 5.8 MBtu/bbl contains 98% Cn H2n+2 ; beloved fuel.
PNGV—Clinton’s Partnership for a New Generation of Vehicles, shared research
between US/Detroit to lower consumption.
present value—future costs and benefits should be reduced to their present values,
a good way to do economics.
PV—photovoltaics crystals of silicon, GaAs or other materials convert solar flux
into electricity.
pumped hydro—US raises water at night to make 20 GWe electricity by day.
PURPA—Public Utility Regulatory Policy Act of 1978 allows alternate electricity
producers to sell to utilities.
quantum tunneling—allows D and T to combine to He in spite of repulsive
potential.
reformers—chemical converter that make hydrogen from methanol or gasoline in
board.
regenerative breaking—on-board generator uses breaking motion of vehicle to
produce electricity.
replacement birth rate—many developed countries are below the 2.1 births/
female, giving a reduced population.
Reynold’s number—(flow velocity)(size)/(viscosity); air infiltration in buildings
is both turbulent and laminar flow.
salt gradient solar—Isreal used the 80◦ C temperature difference in brined ponds
to make small amounts of power.
P1: GFZ
482 F. Glossary
tax credit—a common tool used by governments to stimulate new industries.

thermal flywheel—nautural thermal flywheel smoothes temperature cycle; adobe
houses bring night coolth into hot days.
thermal storage power—night time ice-making and cooling water can reduce peak
power needs by 30–50%.
thermolcline—mountain lakes with summertime 20◦ C temperature differences
could make a very small contribution.
thermodynamic efficiency—first law, net-work/heat input; second law, minimum
task-work/actual task-work.
vehical miles traveled—US light vehicles traveled 2409 billion miles in 2001.
Verhulst equation—King Hubbert used a finite resource in the verhulst equation
to modify oil-use exponential growth.
P1: GFZ
G. Index
G.1 Chapters 1–5 on National Security

ABM, 55 composite spread function, 84
ABM demarcation, 71 convolution theorem, 84
countermeasures, 56 Fourier addition theorem, 85
kinetic-kill vehicles, 70 diminishing returns, 101
layered defense, 74 DOE, 437
national missile defense, 56 downblend uranium, 122
SDI listed below, 484 download warheads, 100
Spartan, 55, 59 drawdown curve, 48
Sprint, 55, 59 duty factor, 65
theater missile defense, 71
adaptive optics, 65, 83 Earth penetrating, 443, 470
antisatellite weapon, 73 effective verification, 78
arms control treaties, 78 Eisenhower, Dwight, 108
Atoms for Peace, 108 EMP, 26, 28
endoatmospheric, 59
bare sphere, 10 exoatmospheric, 59
Baruch, Bernard, 105
bhangmeter, 92 Fermi, Enrico, xiii, book
Big Mike, 18 cover
boost-phase, 67 fission, 6
boosted primary, 14 fragment, 6
breakout from START, 48 energy, 5, 11, 15
breeder reactor, 124, 422 fluence, 18, 69
burn time, 12, 33 flux, 18
Bush, George W., 79, 80 fratricide, 42
fuel cycles, 124
conventional conflicts, 50 fusion, 15, 333
cookie cutter smeared, 53 energy, 15
Costa Bravo, 24
countermeasures, 56 GPS, 36
coupled explosion, 93 gravitational bias, 35
critical mass, 8
CTBT, 45, 90 HEU, 121
Hiroshima, 10, 164
denatured uranium, 166, 122 hydrogen secondary, 15
digital image processing, 84 lithium deuteride, 18
483
P1: GFZ
484 G. Index
ICBMs vs. SLBMs, 49 nuclear weapon free zone, 81

implosion, 12 NWS/NNWS, 111
irradiance, 58
off-alert, 41
Joloit, Frederick, 3 Oklo reactor, 13
optical reconnaissance, 81
Kennedy, John, 105 adaptive optics, 83
kill probability, 38 charge-coupled device, 82
CEP vs H, 39 digital image processing, 84
fratricide, 42 infrared, 86
rate of change, 40 nuclear power monitoring, 88
reliability vs. SSKP, 40 optical, 82
SSKP, 38 resolution, 82
two and n shot, 41
kinetic kill vehicle, 70 plutonium, 14, 124, 185, 189, 422
CIVEX, 128
Lanchester equation, 50 MOX, 125
layered defense, 56 production, 124
lithium deuteride, 18 research reactors, 127
LEU, 116, 120 types, 14, 124
240 Pu, 13
Los Alamos Primer, 4
weapons, 12, 124
Minuteman vulnerability, 47 primary, 10
MIRV, 44
Missile technology control regime, quick from the dead, 105
128
Scud, 130 radar reconnaissance, 88
ballistic missile coefficient, 89
neutron large phased-array, 89
diffusion, 12 synthetic aperture, 88
generations, 12 reactor-grade Pu, 124
spontaneous, 13 Richardson equation, 51
nuclear nonproliferation, 5, 105 rocket equation, 31
NPT, 111
non nuclear weapons state, 111 SALT/START/SORT, 79
nonproliferation policy, 111 secondary stage, 14
nuclear weapons states, 111 radiation compression, 17
preemptive counterproliferation, 115 Serber, Robert, 4, 10
former USSR, 122 SDI, 56
September 11, 2001, 11, 114 adaptive optics, 65
special nuclear material, 112, 116, 124 adaptive preferential, 62
nuclear testing, 90 boost phase time line, 68
atmospheric tests, 92 free-electron laser, 66
bhangmeter, 92 laser beam, 62
cavity cheating, 96 laser in orbit, 63
CORRTEX, 99 laser on Earth, 65
CTBT, 79 particle beam, 60
LTBT, 79 rail gun, 74
seismic, 93 sine qua non, 74
space tests, 90 target interactions, 57
TTBT, 79, 97 x-ray laser, 67
underground tests, 93 SNM, 112, 116, 124
nuclear winter, 27 specific impulse, 32
P1: GFZ
G. Index 485
SS-18, 32 gravitational, 117

strategic weapons, 79 gaseous centrifuge, 117
superhardened silo, 39 laser, 119
SWU, 119 separated work unit, 119
Szilard, Leo, 3 US Enrichment Corporation, 122
theater ballistic missile, 72 verification, 77

throw-wt/launch-wt, 33 context, 77
trajectories, 31 effective verification, 78
CEP, 38 high confidence, 100
errors, 34 how much is enough, 99
flat Earth, 33 treaties, 78
GPS, 36 V2 rocket, 32
gravitational bias, 36
non-spherical Earth, 35 warhead monitoring, 113
Runge Kutta, 34 weapon-grade HEU/Pu, 112, 124,
spherical Earth, 33 166
triad, 47 weapon effects, 20
tritium, 15 cratering, 27
production, 16 earth penetrating, 443, 470
overpressure, 21
uranium enrichment, 116 EMP, 26, 28
aerodynamic nozzle/helicon, 119 weapon efficiency, 11
electromagnetic, 118 weapon scaling, 10
gaseous diffusion, 116 weapon yield, 11, 39
G.2 Chapters 6–9 on Environment

air pollution, 137 carbon in atmosphere
acid rain, 138 See atmosphere and climate
allowance trading, 139 production, 204
plumes, 144 projections, 203
scaling laws, 141 sequester, 225
time dependence, 142 sink, 204
albedo, see atmosphere tax, 226
anthropogenic, 204 case/control, 241
association, see epidemiology chemistry, 156
atmosphere chlorofluorocarbon, see ozone
albedo, 206 climate
atmosphere temperatures, AGU, 199
205 See atmosphere, carbon
carbon/temperature link, 213 carbon/temperature link, 213
feedback, 215 general circulation model, 213
lapse rate, 211 Global Change Research Program, 198
latent heat flux, 210 IPCC, 197
multiple scattering, 211 Kyoto, 227
radiation forcing, 214 oceans
Venus temperature, 210 abrupt change, 219
auto emissions, 148 acoustical tomography, 220
currents, 219
BEIR, radiation dose, 168 rise, 217
P1: GFZ
486 G. Index
oceans (Continued) LA basin air, 148

policy, 224 lapse rate, 211
solar, 216 loss of coolant accident, 169
Tombora, 215 light water reactors, 171
confounding bias, 240
magnetosomes, 244
Debye temperature, 171 Milankovitch cycle, 199
Dobson unit, 150 Montreal Ozone Protocol, 150
deposition velocity, 174, 178
National Ambient Atmosphere Quality
ECCS, 169, 184 Standards, 139
ELF-EMF and powerlines, 233 Nuclear Regulatory Commission,
See epidemiology 169
B field, 235
B squared dependence, 244 18 O/16 O, 2 H/1 H,
199
cell phones, 244 ocean acoustical tomography, 220
cyclotron radius, 244 ozone, 150
E field in biomatter, 236
E field, internal fluctuations, 237 particulate matter, 139
E field from walking/rotating, 237 person-sievert, 164
healing, 244 pH acidity, 138
magnetosomes, 244 plumes
Wertheimer-Leeper, 233 137 Cs, 173
epidemiology, 238 reflected plume, 147

case-control study, 241 Pu, 177
expected values, 241 SO2 , 144
Hill’s criteria, 238 probabilistic risk analysis, 181
p and σ values, 242 fault trees, 186
relative risk, 239 parallel/series, 183
Wertheimer-Leeper, 233
Rasmussen, 184
Fick’s law, 144 radiation dose 163
background, 167
general circulation model, 213 BEIR/ICRP, 168
geological repository, 185 full body, 169
funding, 187 radiative forcing, 214
heat load, 188 units, 164
radiation migration, 189 radon, 190
grandfathered, 140 cancer rates, 192
daughters, 190
heat island, 221 infiltration, 191
heavy water reactors, 193 radon levels, 192
HE/IHE explosive, 177 reactor accidents, 169, 181
Hill’s criteria, 238 LOCA in HTGR, 172
high temperature gas reactor, 172 LOCA in LWR, 171, 173
loss of power, 173
Intergovernmental Panel on Climate passive safe, 182
Change, 197 probabilistic risk, 181
International Comm. Radiological radioactivity, 170
Protection, 168 vs. coal, 176, 182
inversion layer, 149, 174, 178 relative risk, 239
relaxation time, 143, 158
Kyoto Protocol, 227 right-of-way, 234
P1: GFZ
G. Index 487
scrubbing, 138 distillation, 155

sequester CO2 , 225 flow, 158
reverse osmosis, 155
Tambora, 215 ultra violet purification,
thermal rise time, 171 154
water, 153 West Antarctic Ice Shelf, 217
G.3 Chapters 10–16 on Energy

balance point, 293 ICF, 333
blower door, 360 indirect cost, 419
Brayton, 273 infiltration, 287
bridging fuel, 258 integrated solar flux, 304
building energy standards, 258, 288
life cycle costs, 414
CAFE, 384 load and skin dominated, 293
capital recovery fee, 368 low-E window, 351
Carnot, 271 LNG, 274
clock thermostat, 294
coal, 267 modal shift, 268
cohort, 276
combined-cycle gas turbine, 273 natural gas, 266
compact fluorescent, 348
constant/current dollars, 406 OTEC, 324
cost of conserved energy, 412
payback period, 411
decibel level reduction, 396 petroleum, 262, 417
degree days, 284 photovoltaics, 317
discounted benefit, 409 PNGV, 379
duty/capacity factor, 423 present value, 411
elasticity, 407 Rankine, 271

elastic/inelastic, 407 regenerative breaking, 388
enthalpy, 271 replacement birth rate, 270
feebates, 386 salt gradient solar, 325

free temperature, 293
Freedom Car, 391 tax credit, 415
fuel cell, 390 thermal flywheel, 310
thermal storage, 345
gasohol, 332, 421 thermocline, 325
thermodynamic cycles,
heat pump, 353 efficiency, 271, 387
high-frequency ballast, 347
hybrid car, 388 verhulst, 264, 269, 417

Hafmeister, Physics of Societal Issues

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Hafmeister, Physics of Societal Issues

Uploaded by

Copyright:

Available Formats

Physics of

Physics of Societal Issues

Cover illustration: Cover artwork by Kathryn Bay, Enrico Fermi photo

Library of Congress Control Number: 2007920188

ISBN-10: 0-387-95560-7 e-ISBN-10: 0-387-68909-5

Printed on acid-free paper.

Preface .............................................................................................. xiii

1 Nuclear Weapons ............................................................................ 3

2 The Offense: Missiles and War Games............................................... 31

3 The Defense: ABM/SDI/BMD/NMD ................................................. 55

3.8 X-ray Laser Pumped with a Nuclear Explosion.............................. 67

4 Verification and Arms Control Treaties............................................... 77

5 Nuclear Proliferation ....................................................................... 105

6 Air and Water Pollution ................................................................... 137

7 Nuclear Pollution ............................................................................ 163

7.4 Weapon Accident Plutonium Plume ........................................... 177

8 Climate Change............................................................................. 197

9 Electromagnetic Fields and Epidemiology ........................................ 233

10 The Energy Situation ..................................................................... 249

11 Energy in Buildings ....................................................................... 279

12 Solar Buildings.............................................................................. 299

13 Renewable Energy ......................................................................... 316

14 Enhanced End-Use Efficiency .......................................................... 343

16 Energy Economics.......................................................................... 402

A. Nuclear Arms Chronology............................................................... 431

B. Energy/Environment Chronology ..................................................... 446

G. Index ............................................................................................ 483

Why This Book?

Einstein is our icon for a no-nonsense, “no-holds-barred” attack on the secrets of

succumb to the temptation of omission. Our goal in applying the techniques of

at Lawrence-Berkeley (Center for Building Sciences), Argonne (Particle Accelerator

1 Nuclear Weapons ............................................................................ 3

2 The Offense: Missiles and War Games............................................... 31

3 The Defense: ABM/SDI/BMD/NMD ................................................. 55

4 Verification and Arms Control Treaties............................................... 77

5 Nuclear Proliferation ....................................................................... 105

1.1 Nuclear Age

1.2. Fission Energetics 5

1.1.1 Nuclear Proliferation

1.2 Fission Energetics

1.2. Fission Energetics 7

1.2.1 Fission Energy

1.3 Scaling Laws and Critical Masses

1.3. Scaling Laws and Critical Masses 9

1.3.1 Nuclear Scaling Laws

n[(η − 1)/τ ](4π Rc 3 /3) = (4π Rc 2 )nv, (1.16)

Rc = 3vτ/(η − 1). (1.17)

Rc = 3vτ /(η − 1) = 3λ(η − 1) = 3/nf σ (η − 1). (1.20)

Figure 1.3. Hiroshima

1.4. Efficiency and Neutron Generations 11

Table 1.1. Critical mass

235 U and 239 Pu critical masses at normal densities are

1.4 Efficiency and Neutron Generations

1.4.2 Folding Paper 51 Times

A paper folded eighty times (corresponding to the fission events in a 15-kton

1.5 Plutonium Implosion Weapons

1.5. Plutonium Implosion Weapons 13

to plutonium (239 Pu):

Table 1.2. Isotopic composition of various grades of plutonium

Pu grade 238 Pu 239 Pu 240 Pu 241 Pu 242 Pu

1.6 Boosted Primaries and H Bombs

1.6. Boosted Primaries and H Bombs 15

1.6.2 Fusion Interaction